web UI

If you have multiple host interfaces it is not very clear which to use for this kind of "ping". Secondly, host not being "pingable" does not mean it is not available via e. g. Zabbix agent.

I think this could be done by allowing additional "Ping" interface where you could specify an IP or host to use for pinging and have global ping interval with default value e. g. 30 seconds.

Availability with multiple interfaces is completely buggy anyway: ZBXNEXT-4677 .
It could be done in a similar way by looking at all ICMP ping items (and flapping if one interface is up, one down...), but it would also need a decision on how to handle multiple ICMP ping items with different parameters on the same interface.
And that's one step short of having custom availability "buckets" for people who want to see HTTP or any other status. And by that time showing triggers there might be an option.

Use models may differ, but if I see SNMP available, for example, it means that SNMP responded to at least one request on the interface defined.  It doesn't mean the zabbix agent or IPMI process is responding (those are different lights).  Just looking at the red/green pattern I have a high level view that is useful.  If I want to know what's really wrong, I'd look to latest data's many detailed diagnostic capabilities.

If my host is responsive on ping but not on ZBX or SNMP (or whatever) it tells me "lights are on" - but maybe not if "anybody's home".   That's fine, it is very helpful to me and I'd think to others. 

External checks are another level.  It is a good point that this might also be a useful indicator and I can definitely see the arguments for it.

I use this host table indicator list in two ways:  setup and monitoring.

During set up I use it to know if the checks I programmed are working or not - which argues either for all possible checks to have an indicator light to show they're working or to have one light for "all OK" (green), "none OK" (red), "some OK" (yellow) 

Maybe hover text to enumerate the checks and status of each (ZBX, SNMP, JMX, IPMI, Ping, external, web, etc. as appropriate).

This is getting more complicated - but when I'm setting up devices, the default configuration is that almost all are responsive to ping and anything else requires configuration on a per-device basis, so ping has some significance at an overview level.

What about push notifications in Chrome and FF?

If it's possible, then yes. Or what about browser sound API? And please make it configurable per user.

pavel, for browser's push notifications there is a ZBXNEXT-3036

It would be nice to have a screenshot here.

Something like this. Could not get the top menu bar to get closer to the Zabbix logo

I can't help but somehow it looks wrong...

Debian 6

That's really big problem for us, we're been maintaining the patch to fix that since 2.2. I'll publish it for versions 2.4, 3.0, 3.2, 3.4 and 4.0 in zabbix-patches repo shortly. See below for result

they are limited to 50 (per template) by default, and you can change this in administration -> general -> gui, "Max count of elements to show inside table cell" option.
does that solve this issue ?

Yes! Thank you, and sorry for the request. I did not see that option until now... is it new?

Could be resolved using a configuration option.

reopen to change resolution.
no, the option is quite old, probably since 1.8.0 or so

Thyis would be an enormous help for everybody trying to use SNMP monitoring with Zabbix!

Yes, I really need it.

Modified version of SNMP Builder for Zabbix 2.4.7: https://www.zabbix.com/forum/showpost.php?p=177909&postcount=498

A better approach would be to have "Add hosts" button or a link in the map configuration form.

Somewhat connected to ZBXNEXT-163

We too had a need for adding multiple hosts to a map. Attached is a patch to do so.  There's a new "Map" tab under Mass Update.  When you choose to link to map, you can then select an editable map, specify a default icon for the hosts, and select the existing element in the map that you want the new hosts linked to.  You can apply the patch by going into your /usr/share/zabbix directory and running the following.

patch -p0 <HostsMassUpdateAddToMap_zabbix3.4.14.PatchFile

Constructive feedback is welcome.HostsMassUpdateAddToMap_zabbix3.4.14.PatchFile

A related feature request for displaying all graphs of a host on a single page: ZBXNEXT-75.

I think that the proposed navigation would make the front-end much easier for users to navigate.

2years ago, and nothing happens?
i think thats are a good idea, because the GUI/GUI-Navigation is not state of the art

any news?

Web Frontend (GUI)

ZBXNEXT-1276 is a bit related

Debian GNU/Linux 8
Linux zabbix02 3.16.0-7-amd64 #1 SMP Debian 3.16.59-1 (2018-10-03) x86_64 GNU/Linux
-----------------------------------------------
PHP 5.6.39-0+deb8u1 (cli) (built: Dec 16 2018 21:56:53)
Copyright (c) 1997-2016 The PHP Group
Zend Engine v2.6.0, Copyright (c) 1998-2016 Zend Technologies
with Zend OPcache v7.0.6-dev, Copyright (c) 1999-2016, by Zend Technologies

Sorry I Miss te version. It's on zabbix 4.0.4. How can I change it?

Hello linuxpizzacats,

We already a ZBXNEXT-2022 for this request, you can check it by following the link and i will link it to this issue.

This issue duplicates ZBXNEXT-2022

Thanks, I´d try to search some issue related without success. I will watch and follow the issue ZBXNEXT-2022.

CentOS 6

this depends on your screen size - just set smaller width for graphs in screen configuration.

zabbix does not do any screen autoscaling at this time.

got it.

Centos 6.0, Apache2

Yes, 3 triggers it's hardcoded limit.
This issue could be resolved in ZBXNEXT-703
But in any case it should be documented.

<martins-v>a note on this limit added at:
http://www.zabbix.com/documentation/1.8/manual/web_interface/configuration?&#graphs

<zalex> that's good, but additional note (see richlv's comment in the ZBX-3193 ) could be added as note too.
Would be great to add it now.

<zalex> currently done in 2.0 doc in three places.
CLOSED

I am closing it.

Debian 6.0.3
mysql Ver 14.14 Distrib 5.1.49, for debian-linux-gnu (x86_64)
PHP 5.3.3-7+squeeze3 with Suhosin-Patch (cli)

When I click one the column "Name", it's fixed.... Sorry, just discovered that right after reporting the bug.
I guess it's related to the recent update to 1.9.9 (beta).

But maybe it's still a bug also for others when upgrading and the old GUI preferences don't fit the new Zabbix version?

I'm sure it has been fixed in ZBX-3366

See https://support.zabbix.com/browse/ZBX-3366?focusedCommentId=51228&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-51228
See http://www.zabbix.com/documentation/2.0/manual/installation/upgrade_notes#upgrade_notes
"After upgrading, sorting in all locations (tables) will be reset to defaults."

I think issue can be closed as Cannot reproduce.

Debian 6.0.3
mysql Ver 14.14 Distrib 5.1.49, for debian-linux-gnu (x86_64)
PHP 5.3.3-7+squeeze3 with Suhosin-Patch (cli)

for the record, couldn't reproduce this with trunk rev 25028 or 1.9.8, both with php 5.3.3 and 5.3.5

I cannot reproduce too in Zabbix trunk rev 25028.
Debian 6.0.3, php5 5.3.9-2

Seems to be user related - when I log in as "Admin", the page displays. But when logging in as another user (Zabbix Super Admin, LDAP authenticated), it fails.

I have the same problem:
PHP Fatal error: Call to a member function setAttribute() on a non-object in /var/www/htdocs/zabbix/include/classes/class.cuiwidget.php on line 101

zabbix version : 1.9.10

I didn't have this problem from the beginning, it started a couple of days back.
I am using only the admin account.

dev branch: svn://svn.zabbix.com/branches/dev/ZBX-4583

svn://svn.zabbix.com/trunk 26616

chrome, firefox

I cannot reproduce in the latest trunk rev 22840.

can not reproduce in trunk rev 22942 with opera 11.52 either

Try Chrome or FF.

Confirm the fix in dev branch ZBX-4315.

Merged to trunk revision 22973. CLOSED.

Chrome browser v14.0.835.202, also tried and have the same issue with Firefox v7.0.1
Both browsers running on Fedora 14

Yes, this problem reproducible in the release of 1.9.6 (beta) ONLY.
But is NOT reproducible in the latest trunk rev 22419 (v1.9.7)

So issue will be closed as "Cannot reproduce"

latest trunk r22162

Fix is available in dev branch ready to test.

Tested

Fixed in revision 22344.

that's terminology project

what we need is zabbix - http://zabbix.org/pootle/projects/Zabbix-2.0/

Is this issue just about renaming the existing text? Rich, can you fix it then?

I suggest to you reserve "S_PORTUGUESE_PT" term for PT (Portugal) and create "S_PORTUGUESE_BR" for PT_BR (Brazilian Portuguese).
This change will be more hard now, but turn easy the PT (Portugal) implementation in the future.

1.9.7v22202 patch - http://pastebin.com/hkZLmcJt

resolved in svn://svn.zabbix.com/branches/dev/ZBX-4208

(1) Not exactly related, but running "make gettext" results in the following error:

$ make gettext
frontends/php/locale/make_mo.sh
frontends/php/locale/de/LC_MESSAGES/frontend.po:6385: a format specification for argument 1 doesn't exist in 'msgstr'
msgfmt: found 1 fatal error

<richlv> incorrect entry will be fixed by translation maintainer; to ensure that such cases are caught sooner, check fuzzy entries while updating po files
<richlv> to clarify, fuzzy entry checking was added in the script; problem with translation has been fixed in pootle, will be merged shortly (but not as part of this branch)

RESOLVED

<asaveljevs> CLOSED

additional changes made : all language strings use direct gettext functions; all language strings use language_region code (also changed in all translations)

trunk rev 22272.

windows locale was not changed yet - determining a need for that was split out as a separate issue ZBXNEXT-994

Reported by Damian Tommasino in https://support.zabbix.com/browse/ZBX-3840?focusedCommentId=43000&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#action_43000.

(1) Please review my changes in r24957. If everything is ok, you can merge.
<Slava>CLOSED

I had another look and realized, that there is a better way to fix it.

REOPENED.

I've implemented a stricter parameter validation. Please review.

RESOLVED.

(2) [GUI]
Please review my changes in rev. 25068.

<pavels> Thanks, CLOSED.

Merged to 1.8 r25074.

The fix for the trunk is available in /branches/dev/ZBX-3955-trunk.

RESOLVED.

Merged to trunk r25132.

CLOSED.

I can confirm this.

confirmed, opera 11.11/win32

It seems to be Opera bug, but i could not find Opera's bugtracker to confirm it 100%.

Couple of links that points on same problem:
http://stackoverflow.com/questions/6084766/changing-button-text-with-javascript-doesnt-work-in-opera-11-11-for-input-typ
http://webinterfacetricks.com/opera_submit_bug/

from the second link :

"You can get around this by also setting its "type" property, even if you set it to "submit" which is what it already was."

would it be reasonable to do that (probably for opera only) ?
if not, feel free to close again, thanks

Opera v10.63 win32 hasn't this problem.
Opera v11.11 has!

This problems is gone in the new version of Opera 11.50 (tested Opera for Linux and Windows too).
Proof on the picture #3 on the ZBX-3315

Close it as it's clear Opera bug.

1.8 branch, rev 19709

hmm, but you might want to add a host and configure everything about before starting to monitor it. i'd say disabled hosts and their entities should be displayed there, otherwise you would be forced to enable host, which might result in alerts being generated etcetc

But this is different than every other part of the system - on Screens, graphs, etc. I don't see old / disabled servers, so it was very odd to see them here; not sure I understand why this particular area would be different.

well, that's just a simple lack of consistency
ideally i guess all selections could have a quickly togglable option to show/hide entities from disabled hosts

ZBXNEXT-317 asks for a similar thing in latest data

(1) [GUI]
All graphs now shown in popup ignoring host/group filters.

<pavels> RESOLVED.

<Vedmak> Now group filter works, but host filter ignored.

<pavels> Fixed. RESOLVED.

<Vedmak> Please review my last commit.

<pavels> Fixed a couple of issues. Please review. RESOLVED.

<Vedmak> RESOLVED

<pavels> CLOSED.

(2) [GUI]
Is the "Status" column in the "Simple graph" pop-up really necessary, considering that we only display active items?

<Vedmak> ye seems it's not needed, i think we can remove it

<pavels> RESOLVED.

<Vedmak> please review my commit 22908, code formatting and inverted negation logic in check if host column should be shown.

<pavels> CLOSED.

Merged to /branches/1.8 revision r22918.

Please review the fix for the trunk in /branches/dev/ZBX-3498-trunk. RESOLVED.

<Vedmak> CLOSED

Merged to trunk, revision 23066. CLOSED.

same in trunk when deleting a host

seems to be fixed in current trunk ?

Cannot reproduce this in trunk or 1.8

Aleksandrs Saveljevs, please confirm that problem no longer persists and we will close the issue then.

Could not reproduce the original issues, but now I get delete-and-clear.png when deleting a template with "Delete and clear".

Sometimes, when I try to delete a trigger/graph and then another trigger/graph, I get an error message about permissions. I do not have a reliable way of reproducing it. Do you have an idea of what could be the problem?

When trying to delete a bunch of host groups, I get cannot-delete-group.png. Apparently, the GUI does not want to delete group "Zabbix servers", but it complains about "Templates" instead.

seems to be fixed in trunk rev 21467.

ok, maybe not. summary :

1. deleting template or host happens ok now;
2. deleting multiple hostgroups seems to show a correct error message now;
3. deleting and clearing an empty template fails with :

Undefined variable: result [templates.php:455]
Undefined variable: result [templates.php:457]

Fixed in dev branch, ready to test.

Fixed in revision 21909.

how could that be done reasonably correctly ?

What I did on my system is mass updating the trappers to their estimated update frequency .
The Required server performance is now much closer to the real vps from internal check Zabbix[cache,value,all]

Thinking about this, one suggestion is to give update interval a real meaning for trapper items , making zabbix frontend aware that a trapper item not updated within 1 or 2 update interval is too old.

That way zabbix could show in overview and elsewhere that the value is not that fresh ... and required performance could be closer to reality

Decided not to include trapped items in calculation of NVPS, it makes no sense since we cannot predict number of values coming.