We recently covered the most desirable feature request in Zabbix – ability to tell an item to collect a value right now, instead of waiting for the next polling interval. But there are many more feature requests – actually, there are almost 2000 open feature requests. Let’s look at TOP20 out of those, and let’s figure out what’s the second most desirable thing to be implemented for Zabbix.

TOP20 Zabbix feature requests, January 2017

Zabbix uses Jira for issue tracking, including bugreports and feature requests. These issues can be voted on, and a large amount of votes on a feature request indicate that majority of users are missing that feature. The leader here is the ability to force item polling, but there are many more. Knowing what those are can be helpful – for example:

• when evaluating Zabbix, as you can easily see what crucial features might be still missing
• these feature requests can also be useful for existing Zabbix users, as the issues often will contain workarounds.

Here’s a list of the 20 most important missing features, according to Zabbix users:

Place	Issue	Description	Reporter	Created	Votes
1	ZBXNEXT-473	Ability to Check Now a specific passive item	Robert Hau	2010-08-17	174
2	ZBXNEXT-927	Allow LLD Graphs to graph an item for each discovered entity in a single graph	Andrew Howell	2011-08-29	150
3	ZBXNEXT-104	switch trigger to OK based on acknowledgement by user	David M. Zendzian	2009-10-23	139
4	ZBXNEXT-46	Host based dependencies	Ankur Sethi	2009-08-09	134
5	ZBXNEXT-97	ability to test an action in frontend	Christoph Haas	2009-10-06	116
6	ZBXNEXT-936	Zabbix Proxy does not proxy remote commands	Jordi Clariana	2009-11-02	115
7	ZBXNEXT-413	Maintenance Period on Trigger/service level	Florian Koch	2010-06-17	100
8	ZBXNEXT-276	LDAP authentication with groups support	Roberto Rios	2010-03-17	94
9	ZBXNEXT-1262	nested host groups	richlv	2012-06-13	93
10	ZBXNEXT-77	Better permission granularity	Johan Segernas	2009-09-15	88
11	ZBXNEXT-18	send an alert after trigger being acknowledged	Sol Arioto	2009-02-09	85
12	ZBXNEXT-117	Add comments to graphs	Robin Holtet	2009-11-04	74
13	ZBXNEXT-1274	Configurable JMX Endpoint	michael	2012-06-15	73
14	ZBXNEXT-75	Add a "show all" option for viewing all graphs for a host on one page"	Jasper Spit	2008-09-29	73
15	ZBXNEXT-247	Built in feature to export reports and statistics	Emil Odepark	2010-02-22	61
16	ZBXNEXT-336	ability to customise inventory fields	richlv	2010-04-29	61
17	ZBXNEXT-10	Need a command-line import/export facility	Marty Shannon	2008-01-14	57
18	ZBXNEXT-158	ability to use discovery returned string as hostname	richlv	2009-12-08	56
19	ZBXNEXT-714	need scalable alternative for the history and items tables	Will Lowe	2011-03-24	53
20	ZBXNEXT-2088	Allow merging of discovered VM's and data from an installed agent	Gene Liverman	2013-12-24	51

While there might be some overlap and duplication between all the open issues, in general Zabbix team does a good job at closing duplicates and cross-referencing related issues, thus the list above should be pretty correct. If you do spot a duplicate, please do speak up 🙂

Creating “summary” LLD graphs

And in the second place we have ZBXNEXT-927, which asks for a way to graph all LLD-discovered entities in a single graph. What is it that cannot be done currently? While LLD allows to create graph for each discovered entity – for example, a graph with incoming and outgoing traffic for each interface – there is no way to track some item across all discovered entities – for example, error count across all discovered interfaces. One could use ad-hoc graphs for one-off cases, or create custom graphs manually, but that is not feasible for anything above a few hosts.

Workaround

While there is no simple and easy workaround, the best way to approach this currently is using the Zabbix API to generate such graphs. Stay tuned, we will soon explore an example script to automatically create graphs with items like these 🙂