Please also see: #ZBXNEXT-8029 for more suggestions about good naming of items.

I agree here

but then we also need a fix for tags in latest data

bc now a filter of lets say 500 will result in items never being shown as maybe there are more then 500 items and there is no visible indication for this even better some tags will probably never show if we filter first on the host group and have selected too many hosts for the end result

Agreed as well.

An item name should describe what the value represents, not where/how the value was fetched (you can check the item type for that).
In the same way a trigger name should describe what the problem is, not where the trigger was originally defined.

Agree having this type of prefix doesn’t make a lot of sense. 

https://www.reddit.com/r/zabbix/comments/174ezc0/template_entities_prefixed_make_so_much_unusable/

https://www.reddit.com/r/zabbix/comments/15xb9kg/windows_linux_template_segregation/

It's an frustrating change, with no benefits for the users, as far as i can see.

Fixed it:

https://github.com/OpensourceICTSolutions/zabbix-official-templates/tree/main

We are planning to introduce better naming in 7.0, stay tuned! I totally agree that the existing naming convention (as is in 7.0 alphas, for example) has excessive redundancy and hard to read.

Highly appreciated alexei, thank you!

Seeing this ticket getting assigned made my heart jump a little, but within a few hours, it became unassigned again ;'(

Will this be fixed in 7.0 or are we going to suffer for another full LTS?

brian.baekel, There are many good reasons to have it implemented and some good reasons to leave it as it is. We are trying to find a middle-ground currently.

I just want to add my voice, that I agree with Brian, that these prefixes are not helpful.
Please remove them, as it has been before.

One of the cases when these prefixes may be useful is when we have a couple of items having the same name but coming from different templates. Like "Nginx: Memory usage" and "MySQL: Memory usage". It is hard to figure out what each item is all about without a prefix.

I have good news! We are removing prefixes in 7.0, coming soon!

YEAH! We appreciate this decision very much! Thank you for listening!

Available in:

• 7.0.0rc1 (master) 547c8cce459

and how should I tell which service is down now?

From which template is the screenshot?

they're all the same now
this one is php-fpm

Emails are also useless now

Subject: Problem: Service is down

Problem started at 15:13:20 on 2024.06.08
Problem name: Service is down
Host: mon-test.******
Severity: High
Operational data: Down (0), 18
Original problem ID: 59

To clarify above comment, someone forget to add the service names back to the "Service is down" triggers 🤓

Please add them back! 😇

Will the someone fix everything?

any workaround? any fix? is someone aware it was a bad idea removing prefixes?
what to do with useless information zabbix 7.0 produces now?

Trigger prefixes will be addressed in ZBX-24866

I believe the decision to remove the prefix was hasty and did not consider its importance in service identification, such as for apache, php-fpm, etc. This change has resulted in all triggers appearing identical.

Regarding tags, they are primarily used to facilitate search, filtering, and indexing within Zabbix. For example, when filtering Apache triggers, tags are utilized rather than the trigger name. However, for clarity and visual identification and notification, the trigger must retain the prefix. Worrying about monitor resolution or trigger quantity is a mistake.

This mistake was mentioned before being removed, but they still decided to remove it.
In my opinion, please revert this change as soon as possible. 

I think the same, these prefixes are VERY important and i glad these rollback on 7.2 version.

The Zabbix interface has become as uninformative as possible. Now we have many items and triggers with the same name, but affecting different services and it is no longer possible to understand where the problem is just by looking at the dashboard. ZBX-24866 won't solve the problem of what to do with the Latest data:

pavmez 

You can use "Show details" and get more informative name.
Also there exists Tags which could be used for filtering and item identification.

This change should have never happened without template- and host-level tags also being shown, searchable and usable in all views and widgets.

Currently the Latest Data overview turned quite useless as pavmez pointed out: You no longer see what application an item belongs to and you get a lot of duplicate item names representing different data. You also can no longer filter for a specific application as by default no item-level tag is indicating the application (this is only indicated by the template-level 'target' tag, but that is not available on this view/level)
(The Show details as morinskis suggested is not something I want to start explaining to non-tech savy Zabbix end-users. They are not interested in the item keys, they want a quick overview of the monitored data)

This change also breaks Data sets based on Item patterns in dashboard graphs, pie''s, etc. With the example Pavel gave: How do you now get the graph of only the number of php-fpm  processes ? This is no longer possible as there are 5 identically named items in the example, hence even when the full item name is specified as pattern, you get a graph with 5 unrelated items. Graphs also don't support filtering on tags, so even if there where item-level tags indicating the service, you still would not be able to create this graph using Item patterns. (Yes this can be solved with a data set based on an item list, but there are other cases where this is not feasible and you lose the power of the Item patterns)

The prefixes removed from the trigger names is also disastrous as you no longer see in alert notifications or on the Problems view which service or application is now down, or which application was restarted, or which process now uses too much cpu etc.. notifications are now useless too due to the missing context. You now always have to go look for details or even go into item/trigger configuration to check what the trigger is actually about.(which is in turn not available to read-only users)

I understand the initial problem with the OS-prefixes "Linux:" and "Windows:", they indeed caused some trouble, but to solve that I think it would have been better to prefix those items with "OS:" like:
OS: Number of running processes
Apache: Number of running processes
PHP-FPM: Number of running processes
...

And maybe there are some other templates that also could use a more generic item-prefix to allow for easier use in widgets and filtering. But certainly not any application specific template. At least not until the template-level tags are visible and usable everywhere (item-level and trigger-level, ..)

So I do concur that the prefix-approach is not the ideal approach but it was the best we had with current Zabbix limitations on tags. Alternatively and in my opinion even more clear would be that the Latest Data and Problems view also have a column showing the originating template-name of the items; that would then also need a macro containing the template name for use in alert notifications as well as search/filter abilities on it in widgets and views. This would allow for people to sort on that column or the name column depending on the current needs or preferences; to filter on application, etc..
But just removing the prefixes without any alternative was an extremely bad idea.

If I have items from several templates that relate to the same component, they now won't group together anymore since they include a prefix.

Now it is the other way around, metrics from the same application are no longer grouped together, but spread across the latest data view. I find it generally much more intuitive and clear that application metrics of the same application are grouped together. We lost that before when the 'Application' was deprecated, then it was sort of re-introduced with the prefixes, but now it is gone again.

The only way the current templates are now usable is by defining a separate 'host' in Zabbix for each template you want to assign. This way the hostname can provide the now missing information. But that is a completely different approach to how, I assume, most people have organized their monitoring; But if that is the now recommended way, then that should be made clear in the documentation..

I totally agree with robinr considerations: the effect of this change is a dramatic loss of usability and comprehensibility.

For example, even a Zabbix experienced admin will not be able to properly interpret the following screen:

These items belongs to the "PHP-FPM by Zabbix Agent" template, but there is no mention in either item names, item descriptions or item tags. Users can also be fooled into thinking that these items represent the number of Linux processes.

The same happens with trigger names like "Service is down" or "Process is not running": how can the user know which service is it?

[Edit: these examples were already mentioned in previous comments. Sorry for duplication. Anyway: a good tradeoff must be found to solve this issue. Current templates implementation is frustrating]

Thank you robinr  for your feedback with detailed explanations. We appreciate your input very much.

Trigger prefixes will be reverted back. Fix is coming soon: ZBX-24866

We are working on these topics to resolve them in upcoming releases:

• Inherited (template, host) tags in Latest data: ZBXNEXT-9546

• Graph widget improvements to filter/display identical items: ZBXNEXT-9547

• General improvements to distinguish identical entities: ZBXNEXT-9327

Trigger prefixes are reverted back (ZBX-24866) starting from

• v7.2.0rc2, release notes https://www.zabbix.com/rn/rn7.2.0rc2 
• v7.0.7.rc1, release notes https://www.zabbix.com/rn/rn7.0.7rc1