This can be done already with the existing functionality.
You need a trapper item and a trigger. No manual entries in the HW inventory field though.

1. Create a trapper item.
2. Make sure you send the item value as UNIX timestamp and display the unit as unixtime
   
3. Create a trigger for this item which takes the unixtime value of the trapper item and compare it to the current time. Then alert if the change is less than your desired time period in seconds. As example 14 days.

   {Template ZBX Test:hw.warranty.last()} - {Template ZBX Test:hw.warranty.now()} < 1209600
   

   

Closing, as solution to the problem is proposed.

If the problem is still present, please, reopen the issue.

Ok, porém de onde você está coletando o vencimento da garantia do Hardware?

Não seria mais fácil, coletar esta informação do inventário?

Em vez de popular o inventário automáticamente, esta informação é populada manualmente, e usa esta informação para gerar alertas no próprio zabbix.

 
Okay, but where are you collecting the warranty expiration from Hardware?

Is not it easier to collect this inventory information?

Instead of populating the inventory automatically, this information is populated manually, and uses this information to generate alerts in zabbix itself.

Zabbix 2.0.6, Mysql server

On Server log:

31628:20130424:181051.289 In substitute_simple_macros() data:'/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: {TRIGGER.NAME.ORIG}"'
 31628:20130424:181051.289 End substitute_simple_macros() data:'/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: name {$TEST1}"'
 31628:20130424:181051.289 In zbx_execute_script()
 31628:20130424:181051.289 In zbx_execute_script_on_agent()
 31628:20130424:181051.289 In substitute_simple_macros() data:'10050'
 31628:20130424:181051.289 In get_value_agent() host:'Zabbix server' addr:'127.0.0.1' key:'system.run["/usr/local/share/zabbix/externalscripts/test1.py \"Trigger name: name {$TEST1}\"","nowait"]'
 31628:20130424:181051.289 Sending [system.run["/usr/local/share/zabbix/externalscripts/test1.py \"Trigger name: name {$TEST1}\"","nowait"]
]
 31628:20130424:181051.291 get value from agent result: '1'
 31628:20130424:181051.291 End of zbx_execute_script_on_agent():SUCCEED
 31628:20130424:181051.291 End of zbx_execute_script():SUCCEED
 31628:20130424:181051.291 In add_command_alert()
 31628:20130424:181051.291 In DCget_nextid() table:'alerts' num:1
 31628:20130424:181051.291 End of DCget_nextid() table:'alerts' [29:29]
 31628:20130424:181051.291 query [txnlev:1] [insert into alerts (alertid,actionid,eventid,clock,message,status,error,alerttype,esc_step) values (29,5,14462,1366805451,'Zabbix server:/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: name {$TEST1}"',1,'',1,1)]

If i run script throw zabbix server

12631:20130424:221255.657 In substitute_simple_macros() data:'/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: {TRIGGER.NAME.ORIG}"'
 12631:20130424:221255.657 End substitute_simple_macros() data:'/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: name {$TEST1}"'
 12631:20130424:221255.657 In zbx_execute_script()
 12631:20130424:221255.657 In zbx_popen() command:'/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: name {$TEST1}"'
 12631:20130424:221255.658 End of zbx_popen():6
 12631:20130424:221255.658 In zbx_waitpid()
 13741:20130424:221255.658 zbx_popen(): executing script
 12631:20130424:221255.668 zbx_waitpid() exited, status:0

If I run script throw zabbix agent
Agent log:

 10379:20130424:181051.289 Processing request.
 10379:20130424:181051.289 Requested [system.run["/usr/local/share/zabbix/externalscripts/test1.py \"Trigger name: name {$TEST1}\"","nowait"]]
 10379:20130424:181051.289 Executing command '/usr/local/share/zabbix/externalscripts/test1.py "Trigger name: name {$TEST1}"'
 10379:20130424:181051.290 Sending back [1]

Actually args in script output(simple script write itself args to file):

['/usr/local/share/zabbix/externalscripts/test1.py', 'Trigger name: name {}'

As you can see macros was lost when zabbix server(agent) execute script.

This problem is related from ZBXNEXT-1643

what's the deal with usermacros {$TEST1} vs {$TEST} ?

Sorry it was typo. Actually {$TEST1}

Andrei,

User Macro equal {$TEST1} -> gantest

This information is related to this problem? I didn't understand it.

Alexander,

Yes it is value of Macros $TEST1.

Additional information

I change name in trigger to {$$TEST1}
and in script output show this

['/usr/local/share/zabbix/externalscripts/test1.py', 'Trigger name: name {26139TEST1}']
As I inderstand it is number of pid.

But, the user macro should remain as is here. I.e. all macros from trigger name should be sent to script without any changes.

I found where was problem.
It need escaping variables with single quote ('), it is not connecting to zabbix.
Finally I use for run script:

/usr/bin/python /usr/local/share/zabbix/externalscripts/test1.py 'Trigger name: {TRIGGER.NAME.ORIG}'

And it is work good

['/usr/local/share/zabbix/externalscripts/test1.py', 'Trigger name: name {$TEST1}']

I'll closing this issue, It isn't bug.

Thanks.

If expression uses in action throw the script, there is a some problem.

action message:
Trigger name: ''{TRIGGER.NAME.ORIG}''

in server log when send message:
9664:20130425:172034.832 In zbx_popen() command:'/usr/local/share/zabbix/alertscripts/test.sh "123" "PROBLEM: name gantest" "Trigger name: name {$TEST1}"'
shell get it how string with variables, {$TEST1} disappeared 

in output 
Trigger name: ''name {}''

I found it in code and do some changes, and It will work:

alerter.c: 103:
-- zbx_snprintf_alloc(&cmd, &cmd_alloc, &cmd_offset, " \"%s\" \"%s\" \"%s\"",
					send_to, subject, message);
++ zbx_snprintf_alloc(&cmd, &cmd_alloc, &cmd_offset, " \'%s\' \'%s\' \'%s\'",
					send_to, subject, message);

after in server log:
27816:20130425:183849.634 In zbx_popen() command:'/usr/local/share/zabbix/alertscripts/test.sh '123' 'PROBLEM: name gantest' 'Trigger name: ''name {$TEST1}''''

and output
Trigger name: name {$TEST1}

I think need more deep testing about this patch.

I have tested shells but it was without result. (zsh,tcsh,ksh)
All cases shell think it string with variables.

what's the current state, is there anything to fix or improve in zabbix ?

Debian Zabbix

I have finally found the issue: It seems that they're disabled until that specific date in august and I cannot change that from the GUI - in fact, I don't even get any info until when the host is disabled! Please include that in the GUI somewhere. The disable_until value should at least be cleared when activating a host.

select host, from_unixtime(disable_until) from hosts;

+----------------------------------------+------------------------------+
| host                                   | from_unixtime(disable_until) |
+----------------------------------------+------------------------------+
| db.rzl                                 | 2013-08-04 07:11:28          |

Yes, should be cleared all hosts.disable_until*, drules.nextcheck, httptest.nextcheck fields in this case. Zabbix server should be restarted.

I'm closing the issue.

(1) i'd hate to lose this bit of info, so it's added at https://www.zabbix.com/documentation/2.2/manual/appendix/faq - please, review

sasha I added "proxy" into an answer.

<richlv> thanks, CLOSED

ZBX-10906 is very similar

while this issue could be closed, maybe it's worth adding some sanity checks ?
i agree that clock should be correct, but the problem is that having incorrect clock just once penalises the user heavily (and without any feedback on why shit doesn't work).

for example, maybe it's possible to reset "disables" is they are more than some time in the future ?

if not, feel free to close this issue

CentOS 6.5

According to https://www.zabbix.com/documentation/3.0/manual/introduction/whatsnew300#multiple_oid_support_in_snmp_discovery
there is new syntax for SNMP OID in discovery rule.

But I could not find in documentation (especially on upgrade 3.0 notes) any information that old syntax (just OID) is not supported.
It even "sounds" more like new syntax is supported additionally to existing one:

SNMP discovery item OID value now accepts a discovery[{#MACRO1}, oid1, {#MACRO2}, oid2, …] format (see low level discovery for details). The OID values of existing SNMP discovery rules will be converted to the new format (discovery[{#SNMPVALUE},<old oid value>]) during database upgrade.

https://www.zabbix.com/documentation/3.0/manual/installation/upgrade_notes_300

Spec says:

Existing SNMP discovery rules will be converted to take advantage of the new syntax, old syntax will no longer will be supported.

https://www.zabbix.org/wiki/Docs/specs/ZBXNEXT-1554

I believe we have to:

• add "old syntax will no longer will be supported."
• change "SNMP discovery item OID value now accepts a discovery" to "SNMP discovery item OID value now uses a new discovery"
  at:
  https://www.zabbix.com/documentation/3.0/manual/installation/upgrade_notes_300
  https://www.zabbix.com/documentation/3.0/manual/introduction/whatsnew300
  and probably to https://www.zabbix.com/documentation/3.0/manual/discovery/low_level_discovery#discovery_of_snmp_oids

(1) also, as we started to use special syntax for OID - why we didn't perform OID syntax check similarly to item key ?

Thank you Oleksiy. All made in the instructions. Began to work. Good modification. Previously, there was no access to the documentation 3.0.
Спасибо.

The OIDs to discover are defined in SNMP OID field in the following format: discovery[{#MACRO1}, oid1, {#MACRO2}, oid2, …,].

sure, but as zalex noted, there are some potential improvements in the documentation and the frontend

Another problem with such message is that value comes before the error and if the value is big enough error gets truncated in the DB and can't be seen in the frontend. Error is typically short and way more important for troubleshooting than the value.

It is better to avoid such error messages from your scripts. Zabbix operates with raw values, you must care about script output.

Dear dotneft, you didn't get the point. Message is:

Item preprocessing step #1 failed: cannot perform regular expression match on value "X" of type "string": pattern does not match

and is produced by Zabbix server. Reporter suggests to limit the printed length of X (which is a raw value before preprocessing).

Sorry, I see, in this case I suppose it is good idea to remove such messages at all. It is really not warning message

Good idea from dotneft: don't include the raw value in the message at all because it can change quite frequently and will cause lots of unnecessary

error reason for "host:item" changed: ...

in the log file and lots of unnecessary updates of items.error in the database.

Limiting the value of X would be brilliant, but as Alexey says I think the error is somewhat irrelevant. The error is saying that the regular expression didn't match anything in the string, which....is fine!

ZBX-2692

this issue would benefit from a change description

ping ? what actually was changed, how could it impact user experience, what possible problems could this introduce ?

dev branch rev 14145.

validation missing for :

items;
triggers;
graphs;
applications;
regular expressions (both name & test string);
maybe possible for hostgroup selector in host/template properties ? (would have to take into account 'new group';
all import pages - file field (including image uploads);
maintenance - maybe host/group maintenance ?
it services;
images;
ldap auth;
installation wizard

what about fields where something has to be selected (hostgroup, user etc) ?

database inconsistencies should be much more rare with 2.0 - it introduced foreign keys and other checks. some are still possible, though (like application-item linkage in nested configuration etc)

Maybe the referenced script might be a reference for checks.

I fully agree, preventing problems is better than discovering problems

dublicated by ZBX-24034

It cannot. There is no way to distinguish between "never received" and "throttled" value.

Looks like duplicate of ZBX-16162. Feel free to re-open if you disagree.

I am interested in calculating the values for triggers.

CentOS: v7.9
Zabbix: v5.0.14
DB: PGSQL (postgresql12-server-12.7-1PGDG.rhel7.x86_64) + TSDB (timescaledb-2-postgresql-12-2.3.1-0.el7.x86_64)

As solution options:
1.tuning postgres for forced use of the index (SET enable_seqscan = OFF)

SET example is a session only and will not help for a running daemon.
https://www.postgresql.org/docs/current/sql-altersystem.html

If your table is big and data distribution is irregular, raising the default_statistics_target may help.

ALTER TABLE ... ALTER COLUMN ... SET STATISTICS X;

https://www.postgresql.org/docs/current/runtime-config-query.html#RUNTIME-CONFIG-QUERY-CONSTANTS

Partially fixed in ZBX-20644.

Why not to use https://www.pgbouncer.org/ ?

pgbouncer being another component that could fail and might be too much of overhead to add extra machines for pgbouncers, setting up VIP between them to direct traffic towards correct pgbouncer, while in that case you already would direct traffic against vip that is assigned to readwrite node. 

The whole point of supporting connectionstring would be to avoid adding extra components in the chain for HA solution.

I would not even go for JDBC connectionstring, as it is coming from Java  side, but since zabbix is made in PHP, then the https://www.php.net/manual/en/function.pg-connect.php function seems to support libpq and it's format https://www.postgresql.org/docs/current/libpq-connect.html

In that case the connectionstring could contain all the hosts in cluster and choose the one cluster that is read-write 
postgresql://host1:5432,host2:5432,host3:5432/zabbix?target_session_attrs=read-write
Unfortunately currently zabbix server puts together connection data from multiple variables and does not support putting multiple DB hosts

https://www.alibabacloud.com/blog/597797 Seems to be example of how it is configured in here.

Since the connection creation towards correct node is done still by the same library, then no extra components will be used and no extra links in the chain will be introduced with that change.

In current setup we have set up vip in front of PG servers and in the servers there is also HAProxy installed just to direct traffic to read-write node. It is ugly and has extra parts that could fail, but it gives the chance that DB has done failover to another node, but for some reason VIP did not follow, then the DB connections still end up at RW node.

With the connectionstring potential usecase for performance optimization could be to direct write requests towards rw node and read requests to any of the node in the whole cluster, could allow scalability on read requests, which zabbix will get more to load the values for graphs. That would probably mean that in config there could be RW connectionstring and RO connectsionstring. If RO one is not filled, RW will be used.
Directing RO requests to RO node has potential flaw of replication lagging, that could hide latest written values in some cases, but lowering lag and sync process will be duties of Postgres DBA setting up the cluster.

I would like add my vote for this request!
Thanks so much

pgBouncer is not solution in here at all. pgPool-II could be possibly used in place, but it is hassle to set up pgPool-II for this purpose and another alternative is to set up HAProxy on zabbix server hosts, that forward traffic to read-write postgres node. The HAProxy solution is currently used for me, but in there the best would be to have Layer 4 proxy mode and do health checks and RW node checks in Layer 4 as well and it is annoying if you happen to have password on the role checking for RW node.

So I took half a day and generated patch for this feature zabbix6_pg_connstring.patch

I tested PHP side and seemed to work fine, even when switchover from one node to another. Was not able to build C++ Server application to test it, since working on Windows mainly and setting up build environment currently would take longer than needed.

This patch seemed to be good for tag '6.0.0' and checked out commit was 5203d2ea7d901cd33d148f20586e2155901a7faa.
PHP side seemed to work fine even for origin/master as well

Patch implements Postgres connectionstring support and support to add target_session_attr on Postgres connections when using old parameters. In the old code it is possible to add multiple hosts on DBHost field, making it work quite similar to what connstring is implementing. But for automatic failover the target_session_attr is required.

Updated patch to current master (6cfe791429d7cccda0205335868976cac6c51039) zabbix6_4_pg_connstring.patch

Patch still has both changes, one to add target_session_attr config option (with default value of "read-write") and other one to use connectionstring instead of all the connection parameters. 

Even the target_session_attr change would give possibility to write multiple hosts on DBHost field and get correct HA capabilities. If you would like to keep old behavior, then target_session_attr could have the default value of "any", but would be good to have possibility to configure it.

This change is vital for Zabbix HA capabilities, without database zabbix is nothing. It is not normal situation to change Zabbix config if for some reason DB should have to do failover to replica node. 

If you still suggest using something like pg_bouncer to do that failover, could you please provide sample configuration for it, cause so far I have not seen pg_bouncer having such possibility to proxy connections to R/W node. 

Since there has been bigger changes, then I am not so sure if new patch contains all the required changes as I still do not have build environment set up to compile zabbix_server and zabbix_proxy.

If more load-balancing from DB is wanted, then going further it would be nice to configure read-only db config (by default being same as rw config) and having queries tagged as read-only or queries that modify data and then directing read-only queries towards config for read only and queries that modify data only towards RW config. That way it would be possible to spread the queries across DB cluster. At least talking from Postgres point of view when hot_standby is used, but I am sure similar setup is possible with MariaDB and Oracle.

I'd very much like to see this feature in.

I haven't tested the patches, but it seems to me that stripping one of the if statements of parentheses would make the translation unit incorrect.

This could also be a good addition if you use pg_auto_failover. Would like to see this as an way to make Zabbix truly HA with the great inbuild server HA + PostgreSQL HA with something modern.

"Implementing client-side High Availability is included in PostgreSQL’s driver libpq from version 10 onward. Using this driver, it is possible to specify multiple host names or IP addresses in the same connection string."

I was in the process of writing a patch to implement this for 7.x, but it appears that there's a work-around good enough for me and perhaps others that doesn't involve patching the sources.

What I needed from my WIP patch was a couple of things:

1. the ability to set multiple hosts and their ports
2. the ability to set target_session_attrs
3. ideally no regression for other DB servers so that the patch could be pushed upstream.

It had already been possible to set multiple hosts in the DBHost field and it worked with Postgres as intended, but not multiple ports which I needed. The problem was that DBPort is converted to an integer and stored in a variable common to all database connection code, so I had a dilemma how to implement what I wanted but at the same time keep everybody else happy.

Someone over in the IRC channel asked whether pgservices could work and I thought that it's a nice idea, but there's a similar one, but simpler: environment variables PGPORT and PGTARGETSESSIONATTRS. This may be surprising, but it appears to work, zabbix-server doesn't do anything about it and libpq handles them as designed. You'd think for consistency I should also set PGHOST or PGHOSTADDR, but if you don't set it in the zabbix-server config, it's going to default to localhost, which in my case beats the purpose of everything else mentioned here.

I think ideally zabbix-server should cease trying to parse and handle all possible connection parameters for all database engines that it supports and instead expose a single configuration setting for the user to provide the connection URI.

In fact, it is strange that there is still no such possibility. It would be great if such an option appears

Thanks @pstef

I opted to use environment variable PGTARGETSESSIONATTRS to solve my problem. In configuration added multiple DB hosts. My setup was easier since used default port, but different hosts. Made similar change in PHP FPM config to pass same environment variable to web interface as well. Been using that method for couple of months in production, so far on switchover, there is minor hiccup as zabbix detects that connections should be recreated on DB switchover, but so far system has been able to behave normally. This method works only because they are not setting it to some value. My patch itself would set some default value, but at the same time give you option to set it in config. Potentially could have added multiple ports option in ports field too, but it was more complicated than to add connectionstring field.

I don't bother updating my patches for it any more, big portion of work is already done in there and it is up to Zabbix guys to either use it and merge it to upstream or not. 

I agree with you, they should just give the option to enter one connection string field and not bother with parsing and handling them all. But for sake of backwards compatibility probably those fields should stay. However still could add connection string field and if it is entered, then not try to parse rest of DB config parameters and trust that user configuring already knows what they are doing. Not all of the users will need such option, but some might. When writing my patch I kept in mind that old way should still work, it should include config file descriptions to my best knowledge and should not break something on other DB engines, but in the end it is zabbix dev team choice to use that patch or not. So far it seems this ticket has not gotten any of their attention and will be among other unsolved bug reports. Yes there are more parameters you could set and parsing them all could be too much, but most commonly used ones are sslmode and target_session_attrs, that is why potentially those should be in config while rest could come from environment variables.

Still there could be improvements made in backend and to make zabbix server and web aware of query types and if they are read-only queries or need to modify data. If it is read-only query, then it could be loadbalanced using prefer-standby for those queries and direct write queries to read-write node. It of course brings danger of reading stale data, but that is more on DBA to handle on cluster side.

Production

Hi,

All trigger fields, except "Enabled", based on trigger prototypes cannot be changed. It can be changed on trigger prototype level. After next discovery iteration all changes from trigger prototype will be transferred to the trigger level.

Same situation with triggers, based on templates.

But I am not able to change that even in trigger prototype level. There are also it is grayed out. Is this the problem because I am using the default trigger prototype from the default windows server service discovery rule ?

When I take a clone of that trigger prototype, The Manual close option I am able to check.

But I am not able to change that even in trigger prototype level. There are also it is grayed out. Is this the problem because I am using the default trigger prototype from the default windows server service discovery rule ?

When I take a clone of that trigger prototype, The Manual close option I am able to check.

RHEL 6/7 64bit

Hello Hector,

Thank you for reporting the issue. You can use custom LLD rules to discover all files and add them to the monitoring. I found a similar case registered under ZBXNEXT-1699, so we will close this ticket and continue the discussion there. Feel free to follow the original case and vote for it.

Regards,
Edgar

If you don't need to dive into contents of these files, you can use vfs.dir.count[]. I think it should be sufficient for this use case.

Thank you everyone for your inputs! Will do more testing noting these inputs.

@edgar.akhmetshin
Just because the example use case is about log files, doesn't make it a duplicate of a discovery feature request. Nothing discussed on that ticket will help when I want to know if there exists a file with a particular name, but don't care what the extension is.

Instead of having a vfs.file.exists for each one, perhaps an item like vfs.files.exists to combine this into one item check such as:

vfs.files.exists[/path1/to/some.log, /path1/to/another.log, /path2/to/more.log, /path3/to/another2.log]

Regardless of whether the list of filenames is a single string parameter (“name1,name2”) or the item just accepts multiple parameters with each as a separate file, it would help reduce the chances of an external script timing out.

Ubuntu 16.04 LTS

version 3.4.9 the problem persists.

Can confirm this on 3.4.9.

Here is how to reproduce it:
1) Create user 'another' and assign it to User type 'Zabbix Admin' which by default has no rights to any host.
2) With Zabbix super administrator create a new map and add 'Zabbix server' host inside it. Sharing type remains as 'Private'
3) Authorize with 'another' user. Ensure the map is not visible.
4) Use Zabbix super administrator and include one image in the same map. Save the map
5) With user 'another' refresh the map page. The map becomes visible. It is possible to go inside and see some images (not including sensitive data).
inserting-image-item-in-private-map-allows-all-zabbix-admins-see-some-content.mp4

Hello Aigars Kadikis, your reproduction of the problem was perfect.
This is exactly the problem, I followed your procedure and the behavior is the same as yours, the permissioning of the elements in the maps works perfectly, but if you insert a single element of type "image" then the permission fails, so any user of type "zabbix administrator" will be able to list and view the maps of other users even if they are" private "and even if they do not have any permission on the elements contained in the map "even though the sensitive data is hidden ".

I can confirm that the problem affects all versions 3.4.x and all alphas of 4.0.
Version 3.2.x in my tests are working ok.

It works as documented.

Admin level users can see private maps regardless of being the owner or belonging to the shared user list.

The way it is, I only see 2 exits:

1º Do not use elements of type "image" in the maps. Well, the permissions are respected.
2. Do not leave any users with "Zabbix admin" profile. Well, the permissions are respected.

I have to choose between the 2.

Step by the same problem as Diego and I do not see exit since if not leave some users as admin they no longer have access to some elements

I moved this issue to ZBXNEXT project. Please vote for this change request!

In include/classes/api/services/CMap.php there are some lines that permit this:

 // Setting PERM_READ permission for maps with at least one image.
 $selement_images = self::getSelements(array_keys($sysmaps_r), SYSMAP_ELEMENT_TYPE_IMAGE);
 self::setMapPermissions($sysmaps_r, $selement_images, [0 => []], $selement_maps);
 self::setHasElements($sysmaps_r, $selement_images);

Any updates?

Hi there, I've reported code lines involved in this issues/feature. This ticket was opened on  2018 Apr 07 ...Any updates?

I wonder about this authorization concept.

When I look at this list of references, I can not believe that large companies agree with this mess, or have spent a lot of money on work arounds.

Example: We have several 24/7 data centers. Some have their own site administrators. To be able to administer their own hosts, they need to be Zabbix Adminstrator. This gives them all rights to maps again. The same issue is with Discovery. I am not able to give a DBA for a single host the rights for scheduled downtime without giving it the right Zabbix Adminstrator.
From an authorization perspective, Zabbix is not suitable for large DCs, in particular multi-tenancy environment.
Map is one of many gaps.

By the wording, it is true that an administrator should see everything, but an application admin should be able to set up maintenance or add a template to his host. That's why many of us are Zabbix administrators and have in this case too many rights.

Hello David,
currently you can't manipulate with global expressions through the API.

Centos 6.5/7.1 x64, PostgreSQL 9.5, Zabbix 2.4.7

This error also appear on debug level=2 with interval from 1 to 10-15 minutes.

Here is a longer part of the log on my system:

 14399:20160302:111456.497 Starting Zabbix Server. Zabbix 2.4.8rc1 (revision {ZABBIX_REVISION}).
 14399:20160302:111456.497 ****** Enabled features ******
 14399:20160302:111456.497 SNMP monitoring:           YES
 14399:20160302:111456.497 IPMI monitoring:           YES
 14399:20160302:111456.497 WEB monitoring:            YES
 14399:20160302:111456.497 VMware monitoring:         YES
 14399:20160302:111456.497 Jabber notifications:      YES
 14399:20160302:111456.497 Ez Texting notifications:  YES
 14399:20160302:111456.497 ODBC:                      YES
 14399:20160302:111456.497 SSH2 support:              YES
 14399:20160302:111456.497 IPv6 support:              YES
 14399:20160302:111456.497 ******************************
 14399:20160302:111456.497 using configuration file: /home/zabbix/zabbix-bin/etc/zabbix_server.conf
 14399:20160302:111456.508 current database version (mandatory/optional): 02040000/02040000
 14399:20160302:111456.508 required mandatory version: 02040000
 14399:20160302:111456.517 server #0 started [main process]
 14403:20160302:111456.517 server #1 started [configuration syncer #1]
 14404:20160302:111456.517 server #2 started [db watchdog #1]
 14405:20160302:111456.517 server #3 started [poller #1]
 14406:20160302:111456.517 server #4 started [unreachable poller #1]
 14408:20160302:111456.518 server #5 started [trapper #1]
 14409:20160302:111456.518 server #6 started [trapper #2]
 14411:20160302:111456.518 server #7 started [trapper #3]
 14413:20160302:111456.518 server #8 started [trapper #4]
 14415:20160302:111456.519 server #9 started [trapper #5]
 14421:20160302:111456.520 server #13 started [timer #1]
 14422:20160302:111456.520 server #14 started [http poller #1]
 14420:20160302:111456.520 server #12 started [housekeeper #1]
 14418:20160302:111456.520 server #11 started [alerter #1]
 14426:20160302:111456.521 server #16 started [history syncer #1]
 14430:20160302:111456.522 server #18 started [ipmi poller #1]
 14433:20160302:111456.522 server #19 started [java poller #1]
 14417:20160302:111456.522 server #10 started [icmp pinger #1]
 14434:20160302:111456.522 server #20 started [proxy poller #1]
 14429:20160302:111456.523 server #17 started [escalator #1]
 14435:20160302:111456.524 server #21 started [self-monitoring #1]
 14423:20160302:111456.539 server #15 started [discoverer #1]
 14429:20160302:111520.531 [Z3005] query failed: [0] result is NULL [select escalationid,actionid,triggerid,eventid,r_eventid,nextcheck,esc_step,status,itemid from escalations order by actionid,triggerid,itemid,escalationid]
 14429:20160302:111520.531 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select escalationid,actionid,triggerid,eventid,r_eventid,nextcheck,esc_step,status,itemid from escalations order by actionid,triggerid,itemid,escalationid]
 14429:20160302:111520.531 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14429:20160302:111520.531 database is down: reconnecting in 10 seconds
 14422:20160302:111521.533 [Z3005] query failed: [0] result is NULL [select h.hostid,h.host,h.name,t.httptestid,t.name,t.variables,t.headers,t.agent,t.authentication,t.http_user,t.http_password,t.http_proxy,t.retries,t.ssl_cert_file,t.ssl_key_file,t.ssl_key_password,t.verify_peer,t.verify_host from httptest t,hosts h where t.hostid=h.hostid and t.nextcheck<=1456910121 and mod(t.httptestid,1)=0 and t.status=0 and h.proxy_hostid is null and h.status=0 and (h.maintenance_status=0 or h.maintenance_type=0)]
 14422:20160302:111521.533 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select h.hostid,h.host,h.name,t.httptestid,t.name,t.variables,t.headers,t.agent,t.authentication,t.http_user,t.http_password,t.http_proxy,t.retries,t.ssl_cert_file,t.ssl_key_file,t.ssl_key_password,t.verify_peer,t.verify_host from httptest t,hosts h where t.hostid=h.hostid and t.nextcheck<=1456910121 and mod(t.httptestid,1)=0 and t.status=0 and h.proxy_hostid is null and h.status=0 and (h.maintenance_status=0 or h.maintenance_type=0)]
 14422:20160302:111521.533 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14422:20160302:111521.533 database is down: reconnecting in 10 seconds
 14426:20160302:111521.556 [Z3005] query failed: [0] result is NULL [begin;]
 14426:20160302:111521.556 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14426:20160302:111521.556 database is down: reconnecting in 10 seconds
 14403:20160302:111526.520 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14403:20160302:111526.520 database is down: reconnecting in 10 seconds
 14418:20160302:111526.528 [Z3005] query failed: [0] result is NULL [select a.alertid,a.mediatypeid,a.sendto,a.subject,a.message,a.status,mt.mediatypeid,mt.type,mt.description,mt.smtp_server,mt.smtp_helo,mt.smtp_email,mt.exec_path,mt.gsm_modem,mt.username,mt.passwd,a.retries from alerts a,media_type mt where a.mediatypeid=mt.mediatypeid and a.status=0 and a.alerttype=0 order by a.alertid]
 14418:20160302:111526.528 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select a.alertid,a.mediatypeid,a.sendto,a.subject,a.message,a.status,mt.mediatypeid,mt.type,mt.description,mt.smtp_server,mt.smtp_helo,mt.smtp_email,mt.exec_path,mt.gsm_modem,mt.username,mt.passwd,a.retries from alerts a,media_type mt where a.mediatypeid=mt.mediatypeid and a.status=0 and a.alerttype=0 order by a.alertid]
 14418:20160302:111526.528 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14418:20160302:111526.528 database is down: reconnecting in 10 seconds
 14429:20160302:111530.531 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14429:20160302:111530.531 database is down: reconnecting in 10 seconds
 14422:20160302:111531.533 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14422:20160302:111531.534 database is down: reconnecting in 10 seconds
 14426:20160302:111531.556 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14426:20160302:111531.556 database is down: reconnecting in 10 seconds
 14403:20160302:111536.520 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14403:20160302:111536.520 database is down: reconnecting in 10 seconds
 14418:20160302:111536.528 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14418:20160302:111536.529 database is down: reconnecting in 10 seconds
 14429:20160302:111540.532 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14429:20160302:111540.532 database is down: reconnecting in 10 seconds
 14422:20160302:111541.534 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14422:20160302:111541.534 database is down: reconnecting in 10 seconds
 14426:20160302:111541.556 [Z3001] connection to database 'zabbix_2_4' failed: [0] could not connect to server: No such file or directory
	Is the server running locally and accepting
	connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?

 14426:20160302:111541.556 database is down: reconnecting in 10 seconds
 14403:20160302:111546.527 database connection re-established
 14418:20160302:111546.532 database connection re-established
 14429:20160302:111550.535 database connection re-established
 14422:20160302:111551.538 database connection re-established
 14426:20160302:111551.560 database connection re-established
 14423:20160302:111556.546 [Z3005] query failed: [0] result is NULL [select distinct r.druleid,r.iprange,r.name,c.dcheckid,r.proxy_hostid from drules r left join dchecks c on c.druleid=r.druleid and c.uniq=1 where r.status=0 and (r.nextcheck<=1456910156 or r.nextcheck>1456910156+r.delay) and mod(r.druleid,1)=0]
 14423:20160302:111556.546 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select distinct r.druleid,r.iprange,r.name,c.dcheckid,r.proxy_hostid from drules r left join dchecks c on c.druleid=r.druleid and c.uniq=1 where r.status=0 and (r.nextcheck<=1456910156 or r.nextcheck>1456910156+r.delay) and mod(r.druleid,1)=0]
 14411:20160302:111558.141 [Z3005] query failed: [0] result is NULL [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14411:20160302:111558.141 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14405:20160302:111558.780 [Z3005] query failed: [0] result is NULL [select hostid,key_,state,evaltype,formula,error,lifetime from items where itemid=23278]
 14405:20160302:111558.780 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,key_,state,evaltype,formula,error,lifetime from items where itemid=23278]
 14421:20160302:111600.534 [Z3005] query failed: [0] result is NULL [begin;]
 14413:20160302:111658.151 [Z3005] query failed: [0] result is NULL [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14413:20160302:111658.151 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14415:20160302:111758.162 [Z3005] query failed: [0] result is NULL [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14415:20160302:111758.162 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14408:20160302:111858.174 [Z3005] query failed: [0] result is NULL [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14408:20160302:111858.174 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14409:20160302:112358.213 [Z3005] query failed: [0] result is NULL [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14409:20160302:112358.213 [Z3005] query failed: [0] PGRES_FATAL_ERROR: [select hostid,status from hosts where host='Zabbix server' and status in (0,1) and flags<>2 and proxy_hostid is null]
 14399:20160302:114444.374 Got signal [signal:15(SIGTERM),sender_pid:15844,sender_uid:1000,reason:0]. Exiting ...
 14399:20160302:114446.378 syncing history data...
 14399:20160302:114446.380 syncing history data done
 14399:20160302:114446.380 syncing trends data...
 14399:20160302:114446.392 syncing trends data done
 14399:20160302:114446.392 Zabbix Server stopped. Zabbix 2.4.8rc1 (revision {ZABBIX_REVISION}).

It can be seen that there are failed queries after database connection was re-established. However, these failed queries probably come from processes which did not notice that the database was down before and were using the old connection descriptor for their query. This looks scary, but does not seem to cause a problem: these failed queries will be repeated immediately after reconnection.

Thanks for reply, Aleksandrs!
I've calmed down

From the log attached by asaveljevs we can see that the fact that database was down and came back was noticed by alerter, escalator, http poller (these are processes which use direct queries instead of caches), configuration syncer and history syncer processes (which communicate with database a lot by design). Processes that do not notice such short database downtime are trappers, pollers, discoverer, timer, ... (which use mostly caches or spend a lot of time simply sleeping) Since Zabbix processes do not share information about database status this situation is more or less expected.

After some discussion we decided to leave things as they are. Some of Zabbix processes communicate to database not very frequently. Between rare queries connection handle may become invalid. Query then fails and Zabbix process tries to reconnect immediately. If the reconnection is successful on the first attempt, no messages about database down/up are logged, just the message about a failed query. Rationale:

• We don't want to add "database is down", "database is up" messages on every reconnect, because this would litter logs and frighten users.
• We don't want to add "successfully reconnected" message only, because it would be like ")" without "(".
• We don't want to get rid of failed query messages, because we believe it's not good to hide such things.

As suggested by richlv moved to feature requests to track user interest in this issue.

Zabbix 2.0.2 running on Debian from unstable repository packages.

This might be easily related to ZBXNEXT-310.

We also would like this additional security level.
We've sysadmins that do maintenance work and need to create maintenance windows in zabbix. These sysadmins don't have enough knowledge of zabbix so it would be best to limit there access. Currently they require read-write access to create maintenance windows in zabbix.

ZBXNEXT-310 is an older request but discribes it even better.

this should be handled in the context of overall permission granularity, closing in favour of ZBXNEXT-77

this might be considered w/o full permission redesign, reopening

This little hack gives an actual meaning to "Read/Write" in the context of a regular user. This might be a good idea, because it allows the implementation without massive changes and may reduce the confusion as of what the heck the "Read/Write" column is for.

http://zabbix.org/wiki/Docs/simplify_ad_hoc_maintenance#POC_implementation

I think it's very important for users to be able to set maintenance periods for various hosts and groups without having to be in the Zabbix Administrators group. In fact, the majority of users I have accessing Zabbix I would prefer not be Zabbix Administrators but have the ability to create maintenance schedules.

I would also like this function, very important for larger installations

Add my vote to the list of 'why doesn't this exist yet?' 5 years and counting....

We're assigning dev-teams their own credentials for notifications etc, but allow them to set their own maintenance without needing us to set it for them would save us a lot of time.
Please allow for a role of 'maintenance-user' which can set maintenance for hosts assigned to his/her group.
This makes for a better to maintain rights management while still allowing users to manage their own hosts.

Any news?

How difficult can it be to add a maintenance role. Would be a nice feature. Is there anyway to work around this? With a script or so?

Would my approach work for you @Kornelis?

@Volker:

Is this script compatible with Zabbix 4.2? And how do i implement this?

I agree this feature would be greatly appreciated!

Thanks

@Volker, is it possible to use patch with 4.4? Also, is there any step-by-step instruction how it can be installed?

Given its age I predict it not to apply cleanly. However, I think that the concept is still sound and just try to do the few things that I did in the patch wherever they belong now.

This appears to work on 6.0.2 and I'm not sure about other versions.

Add read-write to the host groups you desire via User Groups and then using User Roles you can restrict configuration of Hosts.

I'm not sure if I'm missing anything, but my users can now add maintenance and appear that they cannot configure anything still.

To make the transition smoother and less painful, I've separated three subtasks that need to be done first: ZBXNEXT-1492, ZBXNEXT-1505 and ZBXNEXT-1491.

subtasks closed, this one can be closed, too

jelisejev Wait, there's still stuff to be done in this task too.

OS Centos 5.7 64bits
DB Postgres 9.0

what's your definition of "inactive" ?

Hosts that are some time without answering Zabbix Server requests. This is the opposite of Auto Discovery.

"Remove host" operation is available for Discovery's actions.
Probably it should help.

as mentioned, if you are using network discovery, you can use host removal operation;
if you want to do that based on triggers only, you can run a script and deal with hosts via the api.

i don't think there's going to be anything more specific implemented -> closing

I updated this feature request summary to reflect a main target - auto-registered hosts automatic removal.

Any nice ideas how it could be nicely implemented are welcome
Just to discuss it.

I'm currently using Auto Registration for load balanced servers I'm starting and stopping with Amazon's AWS auto scaling feature. So what would be useful is an operation to Disable Host after xx mins, then 'Delete Host' after xxx mins

For example when the server is terminated it could be disabled after 60 mins, then deleted after 240 mins.

@Rob, If you have found a way to deal with that type of issues please respond since I'm looking for a way to automatically remove hosts that are no longer available from zabbix.

Thank you

@Ionut Cadariu, I've rencently write a script with python to delete unactive hosts periodly. Call for zbx json-rpc with groupid, we get the status of hosts, then we can decide disable them after specific time and delete them. My zabbix version is 2.2.4, if you're interested, post an email:[email protected] to me.

I'm using zabbix_sender called from the zabbix_agent init script upon shutdown to trigger a server-side API call which disables the host. Then a server-side script runs periodically and deletes disabled hosts greater than X days old through the API. It's working for us but I'd like to see this functionality built-in as ephemeral hosts is becoming more widespread.

@James Hammett

What is the syntax to send the disable signal for zabbix_sender??? I've been searching Google for the past 1.5 hours trying to find exactly that key and value and the best I've found is your comment saying you've done it...

PLEASE reply. PLEASE.

M.

Google is no help as I haven't seen anyone else tackle this problem. I created an item named "disable-host" and I have zabbix_sender send the name of the host to it upon shutdown. I created a trigger that triggers upon each update of "disable-host". That triggers an action that runs a server-side script that hits the API disabling the host. Lots of moving parts but 100% reliable so far. Right now I'm working toward making this all work with Docker containers :^)

please consider using https://www.zabbix.org/wiki/Getting_help for discussion, and maybe even writing a howto on zabbix.org

@JamesH

Man, thank you. I am totally following what you are saying and am working on a Proof of Concept. I had gotten another idea working just late yesterday.

1) Wrote a "service": /etc/init.d/zabbix-disable-host
2) It phones home:

• An "enable" at system boot
• A "disable" at system shutdown
• (also handles "status" which just queries the Zabbix front end for current status)
  3) The back end of it is a python script on the node itself.

I decided on this way because if the node crashes and burns, it never sends the "disable" back so it will go unreachable (and in turn sound the alert)

Currently, I'm trying to figure out how to get the date/time out of Zabbix when the node went disabled. This way we can retain usage history for the node for X days, then delete it. We don't want it deleted right away. Looking at the API docs, I don't see that Zabbix actually stores anywhere WHEN the node got / went disabled. :-\

M.

As a small follow up, it doesn't seem that Zabbix records the date/time a node was disabled.

So what I do now in my "service" file (/etc/init.d/zabbix-disable-host) on stop:

• Sends the "disable (this) host" to Zabbix

  zapi.host.update(hostid=HOSTID, status=new_status)

• Enables inventory in manual mode (so it doesn't try to populate it in "automatic mode" when we don't care)

  zapi.host.update(hostid=HOSTID,inventory_mode=0)

• Writes an epoch stamp to "date_hw_decomm" (Date HW decommissioned)

  zapi.host.update(hostid=HOSTID,inventory={'date_hw_decomm': EPOCH})

Server side I scheduled a daily script to:

• Check that specific inventory key of all disabled hosts

  for h in zapi.host.getobjects(status=1):
     host_id = h['hostid']
     result = zapi.host.get(
       withInventory=1,
       output=["date_hw_decomm"],
       hostids=host_id,
       selectInventory=["date_hw_decomm"])
  
  (after a little monkeying with the json back from Zabbix) you get the epoch stamp recorded when the node was disabled above.
  
  Whip a little math around comparing that stamp to "now" and get a "diff" (number days between "now" and "disabled date").
  
  

• Whack hosts that were "decommissioned" greater than X days ago.

M.

might be solvable by ZBXNEXT-1387

There should be a reliable mechanism (timeout?) to cancel or prevent expensive API requests.
Currently one can just restart the HTTP server. Restarting the Zabbix server is certainly no option.

Otherwise issues like ZBX-6763 might become a serious problem.

Btw, how about implementing this as a separate process accessing the shared memory of the server?

Yes, the current plan is to implement it as separate process(es). This will allow to partly re-use existing functionaly for some requests (for example template linking) and also will give access to configuration/value caches.

Performance issue for template related operations discussed in ZBX-6118

We've had a number of comments from our developers that are creating micro services (which we'll run in application containers) that are looking for a restful end point to send metrics to. Will this be available on the proxy as well and will the API be expanded to allow updating item values?

Can we make this a separate component from the Zabbix server entirely so that it can be installed on a separate machine from the server? The reason I ask is this: the server already has a lot to do, ingesting data from the proxies, writing it to the database, monitoring anything not monitored by the proxies, calculating triggers and items, and running actions based on those triggers in addition to the other various tasks it has to do.

In a large enterprise environment, being able to horizontally scale is a requirement and thus far, we've been able to do that by adding proxies and additional web servers, but there can only be one Zabbix server. We are planning to start using the Zabbix API much more heavily in the next year, so being able to set up multiple API servers behind a load balancer would be extremely beneficial and take some load off the Zabbix server.

I agree with Ross. We definitely need to start separating the server into discrete, horizontally scalable micro-services. May need to look at memcached/redis for distributing in-memory caches between components.

I did consider prototyping a RESTful API written in C but it lead to another million new cool ideas and I couldn't decide where to start. For example, it really should sit in front of a shared Data Access Layer (or two: one for config, one for metrics) which could also be a discrete component with a RESTful/other API and then be used by the front-end, server, public API, etc.

So in summary, please don't move the API to the server binary; let's make it discrete and scalable.

Ryan & Ross, I totally agree with you. The API component must be as independent as possible and ideally shouldn't affect runtime processing of Zabbix Server. We may also give users a choice: run it as part of Zabbix Server or as a standalone process for those who prefer scalability. In both cases the API component shouldn't tightly coupled with Zabbix Server, there will be no shared resources at all.

From the other hand, having API on Zabbix Server side would allow interesting things like building distributed or HA Zabbix Server on top of the API or making all communications (server-agent, proxy-server) to be based entirely on API calls. Just imagine Zabbix Agent doing history.put or Proxy doing config.get. I like it, do you?

when decoupling the api, what about the server caches (value, configuration) ?
those would be very valuable to reuse for the api.
separate caching engines like redis could be considered, but the complexity of a zabbix deployment would increase massively

Yeah I think the caches are valuable to the API but do agree that redis/memcached/other would add install complexity.
Alexei, I definitely like it.

I understand this idea is turning into a complete redesign (Zabbix v4?), but if you will indulge me, I'd like to regurgitate the "million new cool ideas" that I feel may actually have value to you. Take it or tweak it or leave it I guess, but I feel some isolation and abstraction would make Zabbix stand out from the crowd.

In brief:

• All shared memory caches are reimplemented as discrete services with private APIs (i.e. Zabbix implementation of redis/memcached/etcd)
• Each worker process (currently discrete PIDs via fork()) migrated to a discrete service with private API (e.g. trapper, pinger, poller, trigger eval, action queue, etc.)
• Two new Data-Access Layer (DAL) private APIs to abstract away the database (one for time series, one for config/other) to enable multiple backends (OpenTSB, MongoDB, etc.) without affecting other components. The config service could apply templates at runtime rather than storing duplicate data to disk (which is then expensive to update)
• Use a message queue to improve API interconnects (pub/sub), atomicity of changes, producer/consumer load balancing, agent comms, delivery and queuing of metrics, actions, etc.
• a watchdog service to make sure each component is started/listening
• Public API is then a discrete service which leverages private APIs and DAL
• Web front-end should exclusively use the public API ("eat your own dog food")

Advantages:

• Improved (limitless?) scalability of each component (deploy as monolithic server or distributed micro-services (containerised?))
• HA per component
• Significantly improved extensibility (new DAL backends or event brokers would be so much easier!)
• Improved topologies across network boundaries
• Eliminate proxies (replace with ZabbixMQ or agent proxies like SCCM)
• faster changes to templates via config API with updates published on the MQ (no bulk updates or syncers to DB)
• Leverage service discovery protocols (e.g. consul)

Should I put together some sort of proposal on the wiki?

Implemented in svn://svn.zabbix.com/branches/dev/ZBXNEXT-2260

(1) how exactly is this being done ? just by dropping the support ?
if so, this has a potential of massively messing up with existing users that are upgrading - they have huge amounts of items, actions and other configuration entities that will be broken now, and there is no way to easily fix them - they either have to find each instance manually in the frontend, or know all tables in the db where this should be changed and do direct sql.

that would be seriously, seriously bad.

sasha Will be introduced an upgrade patch. All deprecated macros will be replaced by supported analogs.

<richlv> HOORAY, thanks
we will need a nice spec, listing all macros and locations where they will be changed

jurism We need a specification that lists database tables and fields where the macros are stored and what macros can be placed in those fields.

jurism Database patch added.

(2) [PS] please review my changes in r44407.

jurism Reviewed. RESOLVED.

similar to ZBXNEXT-735

mma, if your idea is different from ZBXNEXT-735, please add it as a comment there.

Closed as duplicate.

I think yes, a bit. My idea is applying an item of a template to a generic interface (linked by interface's name). It's not exactly the same idea with ZBXNEXT-735 where justin phelps want assign a template directly on an interface. I think the goal is the same but the way to solve our need is different.

mma, what about: [...] having an optional alias per host interface. This could then be set additionally in template items [...]?

My idea is more flexible but more difficult to implement I think, and more difficult to understand for the end-user.
You can have a template with many items witch have different interfaces assigned. Many templates has standard interfaces (WAN, LAN, DMZ) on Cisco Router or Gateway.

okkuv9xh, yes I haven't see with comment. Our points of view are similar. So you can close as duplicate this request.

ZBXNEXT-735 was implemented in a limited way that current request is not a duplicate anymore.
So reopening and updating summary.

All platforms that have libcurl enabled.

would be best implemented as part of ZBXNEXT-173

ZBXNEXT-173 did not seem to have implemented it, see ZBXNEXT-3554.

I'm asking to raise CURLOPT_MAXREDIRS to 50, as it done in curl upstream.

Hi Guys,

as is usually we rise the exception... using the HTTP Agent with some EMC Unity RestAPI we need to follow the redirect more than 10 levels..

Please before is to late for Zabbix v.5.0, can you extende the limit to 50 (CURLOPT_MAXREDIRS)?

Thanks so much

In Inherited and template macros, create Macro with Effective value for redefine the number of days without installing updates

It sounds like a great idea, let us check if it can be implemented without use of system.run.

Just a quick status update. We are about to start development of this functionality, aiming for Zabbix 6.2.

Development of new metrics (without template changes) for monitoring of Windows registry was moved to ZBXNEXT-7732.

Monitoring of Windows updates using registry seems not to provide real latest info about updates. Sources other than registry should be considered.
Some categories of updates are being installed, but they do not change a timestamp in registry (for example, KBxxxxx updates for Windows Defender, UWP apps, drivers). In such cases eventlog will contain records about successfully installed updates (with event ID 19, and source as WindowsUpdateClient), but timestamp in registry will not be changed. Judging by this fact, Eventlog can be the right source, as WMI behaves similarly to registry. However, eventlog monitoring is available only as active check.

Windows registry monitoring metrics are available in:

• pre-6.2.0rc1 - a38b1cecd97

While implementing registry monitoring was mentioned above this is not what was asked for (Windows Update monitoring). Keeping this open.

Zabbix 6.0 LTS

I do confirm. Replicated on Zabbix appliance 6.0.15 LTS(https://www.zabbix.com/rn/rn6.0.15). Created: user "Admina" with Admin role and group Zabbix administrators; user "User" with User role and group Users(changed during and for the process to Admin role).



User can edit the map.

Created and owned by Admin role type user "Admina".

User can not edit the dashboard.

Created and owned by Admin role type user "Admina".

This works according to the documentation:

Admin level users can see private maps regardless of being the owner or belonging to the shared user list. Private maps can be edited by the owner of the map and in case a user/ user group has read-write permissions for this map and at least read permissions to all elements of the corresponding map including triggers in the links.

Java

Yes, it would be great!

yes , I really need this feature.

JMX operation support would be a nice feature, but it is not that simple...
Operations could have multiple params with different types (not only types from java.lang, but any custom type), so Zabbix user will have to provide a value and a type for params of operations. It is also hard to tell what syntax can be used for arrays, custom objects, etc to define all that within single field and even for fragmented input (type dropdown and input for a value) there are no good options to define objects, arrays, etc.

Provided solution is limited to strings only and uses item key as a data buffer for operation name and param names. Currently item key is limited to 255 characters and adding all the values with their types into one input field will be a nightmare for a user (so fragmented input is a must).

So looks like limited solution (for basic java.lang types) is the closest Zabbix can get to JMX operation support without asking user to write custom scripts or do some magic inducing serialization of params.

It might be worth implementing operation support with plain strings as the first shot. It would allow to verify the basics and additional parameter support can come at a later time.

I run into this problem while trying to monitor the default springboot mbeans. I feel like this is a big limitation of the jmx gateway.

Closing as duplicate of ZBXNEXT-927

How is this a duplicate? That one talks about same host, this one spans across multiple hosts and has nothing to do with LLD. This request is purely to add a feature to create screen with graph from template. And then screen automatically figure out which all hosts using that graph and render them.

Oh, sorry. Misunderstood the description.

Somewhat similar issue ZBXNEXT-927 - asks for grouping graphs of items created from LLD. 

Any plan so far to get this implemented anytime soon? We are migrating to latest version (4.4) and creating all these screens is very tedious without this proposed solution.

Actual for master items that collect prometheus metrics.

Just as another example, kubernetes kube-state-metrics could easily be more than > 64KB

We are going to increase this limit in Zabbix 4.4, stay tuned!

Wow, so fast! Thank!

It seems I misunderstood this feature request. I thought it was about ability to process larger values. I am afraid we won't be able to increase size of the text fields in the DB in Zabbix 4.4. Perhaps later.

Sorry for the confusion.

I thought that limiting the database also affects the processing of dependent data elements. Now I checked on version 4.2 and saw that the data outside 64Kb is being processed normally.

trunk rev 8107

This is a global problem of checking passed data, and is part of future API development what is not covered by a single issue... =) for now it's stays on clients side responsibility.

valid as a feature request (so that it doesn't get lost)

RedHat CentOS 5

Just follow standard NET-SNMP documentation in order to import new MIBs. It is extremely straight-forward procedure.

Alexei

I believe the OP was asking for a way to import a MIB into Zabbix and have Zabbix create a Host Template with all the Items created as defined in the vendor MIB (or allow the user to pick and choose before the template is generated). This would be preferable to building the Host Template by hand in an environment where a lot of different devices are used.

I can't see where importing the MIB file into NET-SNMP really has any effect on Zabbix other than being able to use the MIB OID object name rather than the numeric OID.

then this is a feature request (somehow i can't find an existing one...) for template builder from snmp mib.

until then, http://www.zabbix.com/wiki/scripts/gentemplate.pl seems to be the most functional script to generate templates for snmp devices in an automated fashion.

The link http://www.zabbix.com/wiki/scripts/gentemplate.pl is broken.
--Stefan

yes, it pointed to the old wiki which was in a phase-out state for a long time - if the script was not migrated to zabbix.org, it was probably too outdated and not too useful at that time

For who are interested on a possible alternative to the old tool to convert MIB file, i would suggest https://github.com/cavaliercoder/mib2zabbix . I think is very great tool!

I also suggest to Zabbix to think about to something similar directly integrated on the main interface

very similar to ZBXNEXT-57

dev branch: ZBXNEXT-315

initial info is shown, any ideas what else can be shown?
getting info about server will require additional development for server side.

hmm... i'd argue that it doesn't make sense to introduce a different page.

this partially duplicates 'status of zabbix'.

so if we are doing this, it would make sense to do it properly. current separate page has the issue of.. being separate, as well as completely visually different from anything else (and i've got enough complaints about user interface inconsistencies and usability sucking in the last few weeks to last for several years...)

so what we should do...

unify config wizard requirements page, "status of zabbix" and this info page in a single page.
"status of zabbix" and sysrequirements partially overlap, because status would show problems only, but not current php config values that frontend requires. so there should be :

• a section of the page giving high level overview just like "status of zabbix" does;
• a section of page giving frontend requirements just like config wizard does (so showing the data always, not just errors);
• low level info just like this new info page does.

and what's important - it all should be visually and functionally consistent. current visual representation leaves a very bad impression.

I got an idea of how you can do version, revision and server futures checking, and I decided to tell about it.
It can be done so that the server during the start, wrote somewhere in a database with all necessary information (version, revision, compilation time, futures, etc.)
Web Interface will only read this information from the database.
Give the necessary information simply on port 10051 I think is not safe.

It would be nice to resolve this request at the same time
https://support.zabbix.com/browse/ZBX-2512

yes, server inserting some runtime info in the db was exactly the idea in ZBXNEXT-57

but it would help to save a lot of time by deciding on a consistent implementation. cleaning it up later can turn out to be near impossible, as we have seen way too many times on other areas.

I am reopening this issue as seems that at this moment we do not plan to implement this feature in v2.0, maybe will be implemented later.

ZBXNEXT-1194 asks for similar functionality on the api level, these both should probably be resolved together

CentOS 5.6

most likely a dupe of ZBXNEXT-968 - if not, please reopen

This is not duplicate, reopening it, as time to time people ask for that.

We really need this feature. It will likely be the dealbreaker that will make us switch to using the grafana stack for monitoring infra and services.

Hey Rashawn_Johnson, what's you use case?

We're trying to monitor the logs of a rails app. Below is an excerpt of the logs

I, [2024-04-03T20:00:25.689375 #1]  INFO -- : [d2ec09f4-e1fd-4b05-a49f-2177b80a012b] Started GET "/" for 159.203.36.31 at 2024-04-03 20:00:25 +0000
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/controllers/application_controller.rb:172:in `set_timezone'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/controllers/application_controller.rb:56:in `execute_around_action'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/controllers/application_controller.rb:60:in `switch_locale'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/controllers/service_cards_controller.rb:57:in `endpoint'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/models/endpoint.rb:90:in `perform_action_as_user'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/models/endpoint.rb:83:in `perform_action'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/models/endpoint.rb:83:in `public_send'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/models/trust_adapter.rb:107:in `perform_post'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/services/http_utils.rb:92:in `perform_post'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/services/http_utils.rb:103:in `make_request'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] app/services/http_utils.rb:124:in `make_call'
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9]
[34ea335d-9627-4fc6-8b1f-957d6e2b46e9] URI::InvalidURIError (bad URI(is not URI?): " http://127.0.0.1:2375/v1.24/containers/json"):
E, [2024-04-03T19:59:27.412915 #1] ERROR -- : [34ea335d-9627-4fc6-8b1f-957d6e2b46e9]
I, [2024-04-03T19:59:27.411767 #1]  INFO -- : [34ea335d-9627-4fc6-8b1f-957d6e2b46e9] Completed 500 Internal Server Error in 22ms (ActiveRecord: 16.4ms | Allocations: 2452)
I, [2024-04-03T19:59:26.614839 #1]  INFO -- : [34ea335d-9627-4fc6-8b1f-957d6e2b46e9] Started GET "/fr/identity/password_reset?%F6%7Bconstructor.constructor(13250300)%7D" for 52.60.149.201 >
I, [2024-04-03T17:06:41.466371 #1]  INFO -- : [8eb439ee-af77-486c-93ee-6c4ee3584e0b] Completed 200 OK in 24ms (Views: 5.5ms | ActiveRecord: 13.5ms | Allocations: 4152)
I, [2024-04-03T17:06:41.466201 #1]  INFO -- : [8eb439ee-af77-486c-93ee-6c4ee3584e0b]   Rendered layout layouts/public.html.erb (Duration: 14.7ms | Allocations: 3289)
I, [2024-04-03T17:06:41.441846 #1]  INFO -- : [8eb439ee-af77-486c-93ee-6c4ee3584e0b]   Parameters: {"{constructor.constructor(13246203)}"=>nil, "locale"=>"fr"}
I, [2024-04-03T17:06:41.441700 #1]  INFO -- : [8eb439ee-af77-486c-93ee-6c4ee3584e0b] Processing by Identity::PasswordResetsController#new as HTML

We need to be able to extract the id of requests with error status (`34ea335d-9627-4fc6-8b1f-957d6e2b46e9` in this excerpt) and the error message information which can be found in subsequent lines after the error is logged (on line starting with 'E,'). Different errors will generate different number of error message lines. The request id and message info will then be used in alert messaging. 

We've tried different regex patterns that work in test but always fail with `pattern match not found` till we saw this comment.

Yup, that syntax [[:space:]] should work. 2 other ways are the (?s) inline modifier or the \s token. 

If you're ok with ingesting the entire log into the item, then problem solved! One of the above multi-line regex approaches combined with regex lookahead should be enough to match the lines. 
{{If you don't want to keep the entire log file in a Zabbix item, then you'll need a different approach. }}

To capture only the lines of interest into the item, you'll likely need to "pre-process" your log with a script. The script will need to keep track of where it is in the file in between runs, and use awk or perl with the range operator to define the start and end regex's. Then use zabbix-sender to feed that into an item. }}{{The trigger then becomes simple - every time a new item value is received, create a problem. Alternatively, could get the data into Zabbix via }}{{vfs.file.contents[] or log[] item. 

I have a pre-processing script that you could use as a starting point, if you're interested I'll post. 

Thank you Matthew.Steeves for these options. We hadn't looked into pre-processing the logs yet so I'm interested in your script. 

hmm. usual suggestion is to avoid using 'status' item and use agent.ping + nodata() instead - that should solve this issue

I know agent.ping + nodata() solution, but the solution don't solve the first problem
the most important problem is the first one

And, I think check of agent availability is important, but agent.ping + nodata() is difficult for Zabbix beginners.
If we recommend agent.ping + nodata() for all user, I think it's good to implement that function by default, without creation of item and trigger.

in latest versions userparameters going down still should result in nodata() trigger for agent.ping. would that solve the problem ?

continuing with slow hosts as usual is not feasible, as they would hamper the overall monitoring - such userparameters should be fixed/changed

Kodai, about "if any user parameter item take a long time, Zabbix server stop monitoring all of other item on same host"
Seems you are partially wrong. See my last debugging ZBX-4284

But this feature request is very interesting overall.

Thank you pointing that out, Oleksiy.

first problem I wrote looks same problem as ZBX-4284. Only one "slow userparameter or check" affect to all of other items on same host. I think it is not good behavior. Zabbix server should only change item status to not supported in this case.

I think heartbeat communication between server and agent (and unreachable host is handled by heartbeat) improve the problem and usability so much.

...Only one "slow userparameter or check" affect to all of other items on same host. I think it is not good behavior. Zabbix server should only change item status to not supported in this case. ...

agreed !!!

i had this problem yet,i found something funny that zabbix server will check that server after aboute 320 minutes again.

Zabbix setting a host as unreachable just because one item is bad, is the most annoying thing. The biggest problem is that it makes Zabbix unreliable, because if one item times out, the whole host gets disabled, so it stops running other items on that host, which means the host is not being monitored properly.

I think the easiest way to solve this problem is to define items which can make a host unreachable, e.g. it could be a check box in the item create/edit page called "Make host unreachable on time out", or some better name. This way e.g. agent.ping could have that checkbox set and only that item would make the host unreachable.

that's a different problem which is tracked at ZBX-4284

Well it's not a different problem, they are related, since both of them would change how we detect if a host is unreachable.

Actually. It appears to be the case only when the macro has been defined on template level with an empty value.
If nothing is defined it does properly "not authenticate" resulting in a 200 status.

In any case this is not a bug.
If you like, reopen and move to ZBXNEXT project.

Closed

moving to feature requests, reopening

VmWare ESXi 5.1 host servers

please use zabbix forums, irc and other channels for community support. this tracker is for bugreports.

I'll rewrite that. I need a compiled or a procedure to install Zabbix Agent in ESX Host. As I'd check in forums there's no way to do it as we deploy the agent.

created by now

reopen for moving

(1) We should get rid of testhelpers instead using ZbxExitException instead of exit() / die()

I've implemented basic fixture-based tests in development branch svn://svn.zabbix.com/branches/dev/ZBXNEXT-686

Basic file-based tests and parameter handling implemented in r47620 according to spec.

File-based tests with complicated scenarios and Respect/Validation validation implemented in r47644.

(2) We need to be able to validate arrays using the same validators as separate values. Something like the code below, but simpler:

response:
  __assert: count(3)
  __keys:
    __assert: sequentialKeys
    __each:
      __assert: int
  __each:
    id:
      __assert: notEmpty:int
    hostname:
      __assert: notEmpty|string|length(5,10)
    templates:
      __assert: array|length(3)
      __each:
        __assert: string|length(0,12)

A similar example with equals support:

response:
  __assert: count(3)
  __keys:
    __assert: sequentialKeys
    __each:
      __assert: int
  __each:
    id:
      __equals: 1
    hostname:
      __equals: Zabbix server
    templates:
      __equals: [100, 101, 102]

andrewtch RESOLVED in r47667. Also, a new validator has been added, 'sequence' (implementing sequentialKeys as above).
Also, I've implemented nice exceptions with paths like "Validation definition is not an array on path "_assert->_keys->_each[0]".
I've also added support of short syntax turning:

hostids:
    _each:
        _assert: int

into:

hostids:
    _each: int

in r47671.

jelisejev CLOSED.

(3) We should NOT use direct header() call, using wrapper instead.

RESOLVED in r47747.

jelisejev CLOSED.

(4) We should avoid all

__DIR__ . '/../../../../../.......'

constructs and make some path resolver (probably, these values should be set inside phpunit config file).

jelisejev WON'T FIX.

(5) We should use symfony/console table helper while displaying table rows in sql assertion instead of json_encode.

jelisejev CLOSED.

A lot of changes to the API tests have been made, so the comments above can be ignored.

In short, branch svn://svn.zabbix.com/branches/dev/ZBXNEXT-686 implements a framework for writing API tests.

Some changes have been made to frontend API client. It now fully implements the JSON RPC protocol. Please test that nothing has been broken in the frontend and the API endpoint.

RESOLVED.

ALL

It is already implemented. You can use "Global regular expression" with "Result is FALSE" condition: https://www.zabbix.com/documentation/3.4/manual/regular_expressions

Alexey, this would be great if global regular expressions would work in "Required string" field for web scenarios. But they don't :/

Brian, web.page.regexp agent item might be something to look into, but they also have limitations. 

https://www.zabbix.com/documentation/4.0/manual/config/items/itemtypes/zabbix_agent

Yes, I agree, It is also nice idea!

ZBXNEXT-3806 is about exactly the same problem but a quick test with pcre negative lookahead did not return the desired result. The expression pattern is still evaluated as true thus the scenario step is successful. Some regex magic though might do the trick, I hope. 

^((?!ZZZZ).)*${noformat}
 

Pattern (?s)(?!.*ZZZZ)^ seems to do the trick.

Gleb, you are spot on as always!

However, if one must ask a top class C developer to solve this configuration request then it can't be considered as a feasible option for everyone. Which makes ZBXNEXT-3806 still valid.

Possible solutions would be additional field for negative match in the scenario step or added support for global regular expressions. 

Perhaps, we can ask martins-v to add this workaround as an example.

This isn't working for me in Ver 3.0... Is this solution for Ver 3.4 ?  After the Regex changes?

Thanks

Yes, it uses PCRE features.

Zabbix server on Ubuntu 16.04 Server, several zabbix proxies on ArchlinuxARM, and lots of Windows agents

What if you execute remote commands on zabbix agent ?

Yes, I do so. Like that:
Executing command 'wscript //B //Nologo \\zabbix\zabbix\usr\vbs\setup.vbs /upgrade'
But in step: Shell.Run ZabbixService.PathName & " -x", 0, True
The agent stops along with the wscript subordinate process.
If I use item system.run[wscript //B //Nologo \\zabbix\zabbix\usr\vbs\setup.vbs /upgrade, nowait] then an independent process wscript is created and the update completes successfully.

As a workaround I suggest to execute remote commands through proxy.

The problem is that now it's not always possible to execute remote commands through a proxy. The solution can be to execute remote commands with "nowait" mode.
The workaround is system.run[commands, nowait] item and enabling/disabling its by Zabbix API.

Hello ju5t,

There is no "out-of-the-box" support for Slack in Zabbix. Integration with Slack is done through the custom script shared in our community by the developer of the script. If you have any suggestions - try to contact him directly.

Regards,
Aleksejs!

Since this is not directly related to the Zabbix - I'm closing this request as Won't Fix

Thank you for understanding.

Regards,
Aleksejs!

I know there isn't out of the box support for Slack yet. We're using a similar script as the one you referenced.

Interactive messages need an endpoint where they can post their data to. I think it is far beyond the scope of a community contributed media script to do this. You would have to setup your own webserver to redirect these requests over an API call to Zabbix or change the core of Zabbix to accept POST requests formatted by Slack and convert them into an API call.

It makes more sense to support this as an official media type in Zabbix.

Design available in development branch svn://svn.zabbix.com/branches/dev/ZBX-10810 .

(1) Similarly to ZBX-10777, we should probably not use clip() in CSS.

PavelA REMOVED in r61105

The new design looks good (actually, it is wonderful!), but it does not currently seem to be applicable very much to lists of things (but can be applied in settings like media settings in a user's profile). For instance, in the host list, we display a confirmation dialog. The checkbox should not change its visual state until the user has confirmed his action. Also, since such pages are currently reloaded, animation is not very useful. It may become useful if we later change to AJAX requests. So, in addition to the current widget, there is a need for a non-animatable status switch that can be clicked and wrapped into a link.

All

see housekeeper table, it holds a list of deleted items and housekeeper deletes all data for those items following different rules. i guess we can close this issue then - please reopen if there's still a need to do something

Kindly note the housekeeper has been running for the same years these orphan items exist. Before posting I double checked to make sure the housekeeper completed its run, changing the number of items to process per run to unlimited. Even when that had completed, these history entries describing events that happened to itemids not in the items table remained. I rebooted a few times, let the housekeeper complete, and yet millions of records referring to item record ids not in the items table remain.

So, closing this item on the basis the housekeeper would take care of it needs further explanation.

See reply above.

in general, it would be best to discuss things like this in one of the options at https://www.zabbix.org/wiki/Getting_help - irc would be a great place - but in any case, you did not specify zabbix version. if it's a recent one and you have upgraded, check that housekeeper is enabled at all in administration -> general -> housekeeping

I do think the housekeeper has been properly enabled and configured. Note the version is also in the shot, 2.2.

Really, just a feature request for an option to delete all references everywhere to items that no longer exist.

as noted, that is supposed to happen already. please use https://www.zabbix.org/wiki/Getting_help to discuss the details, until a specific bug or feature request can be reported.

Also as noted, though it was supposed to happen, it was not happening. So here we are on this page where I think it's created to ask for features.

To be clear: The feature requested is a server command line one-of recovery option to delete every orphan item in the database and generally not proceed until the database is in a known consistent state.

Having done this manually, finding a few thousand orphan records in trends* and several million orphan records in history* not mentioned in houskeeper and also not in items I think I've presented a reasonable use case. Look at it this way, if the server flag finds nothing to do it can act as an integrity check and so give confidence the server is about to begin operations on a known-good platform.

It's a reasonable enough feature request IMHO.

Not absolutely sure, but there can be other use cases.
For example we delete a child node (which passed its history to master) on master.
As I recall it was discussed already in other issues.

All sub-tasks done. Review required.

Please assign linked issues for content review.