Zabbix Dashboard applet for items in Not Supported state.

[cdslaughter]

Anyone know of a way to get an applet on the zabbix dashboard that contains a list of host/Items that are in a not supported state?
With over 1600 hosts and 265,000 items, having a dashboard listing (enabled) items/discovery) that are currently showing not supported would be very helpful.

Carl

[kloczek]

No way on dashboards, but why not use Filter in Configuration > Hosts?
Just switch "State" to "not supported" and clear other options or choose interesting groups, app etc.

Of course it is possible to have such data in dashboard
Heree is the screenshot of the fragment of the dashboard which I have now

[kloczek]

Here is the instruction how to do that:
- map to the host item zabbix[host,,items_unsupported] key
- map that item to inventory entry
- go to inventory entry to have URL of the page with agregated count of the hosts with exact number of unsupported items
- add that URL (without sessionid) to the dashboard "URL" widget

My OS templates and SNMP devices templates from https://github.com/kloczek/zabbix-templates have zabbix[host,,items_unsupported] OOTB and that item is always mapped to "Site Notes" inventory entry.
Posted screenshot shows in first column how many unsupported items have some exact number of hosts listed in second column.

[cdslaughter]

Thank you Kloczek.
I will take a look and see if I can make that work for my use case.

Carl

[kloczek]

I've "invented" that trick after submitted the patch with per host counter of unsupported items :P
Top point on my "Zabbix Wish List" is to provide by zabbix custom inventory to be able to define something like "unsupported items" inventory and map to that host item with zabbix[host,,items_unsupported] key.
I've posted even to or three times on the forum what exactly needs to be done to transform existing fixed inventory to fully flexible.
Here is my recent comment about what needs to be done https://www.zabbix.com/forum/zabbix-...493#post384493

IMO if zabbix would be able to provide such custom inventory with only minute API call adjustment zabbix can be used as main inventory for example with any CMDB effectively wiping whole class of that kind of software solution from whole market

[kloczek]

Another thing which still bothers me is extending zabbix which is related to that trick is to provide new type of the metrics which will vector of the data.

Two examples:
1) when for example unsupported items data are mapped to the inventory sample on regular base vector of the data with pairs of number of unsupported items and count of hosts.
Such series f data could be used to drawing 3D graph with x (depth) with number of unsupported items, y (height) with number of hosts with exact unsupported items and z (horizontal axe) time scale.
Generally with custom inventory and vector metrics it would kind of backbone to produce association tables data used in such metrics.

With that it would be possible to spot some disruption of the monitoring causing increase of unsupported items

2) Using DTrace or bpftrace is possible to produce distribution of some values.
Example:

# dtrace -qn 'syscall::write:entry /execname == "mysqld"/ {self->stime = timestamp;} syscall::write:return /self->stime != 0/ {@LWrite = quantize(timestamp - self->stime);} tick-10s {printa(@LWrite);}'
           value  ------------- Distribution ------------- count
             256 |                                         0
             512 |                                         10
            1024 |                                         15544
            2048 |                                         28204
            4096 |@@@@@@@@@@                               713511
            8192 |@@@@@@@@@@@@@@@@@@@@@@@@@@               1829254
           16384 |@@                                       109075
           32768 |@                                        59445
           65536 |                                         9685
          131072 |                                         2268
          262144 |                                         1408
          524288 |                                         621
         1048576 |                                         830
         2097152 |                                         338
         4194304 |                                         377
         8388608 |                                         598
        16777216 |                                         260
        33554432 |                                         9
        67108864 |                                         3
       134217728 |                                         0

Above it is ascii art graph showing distribution of the latencies of the IO operations caused by mysqld.

Other example could be produce distribution of the SSDs latencies to observe how it evolves in time with changing workload or with degradation of the flash memory cells to raise alarm that some maximum of the distribution of the letecies is moving to far.

With vector type of the metrics would be possible to observe how is mowing hump with maximum of the latency of that IOs.

Vector data can be stored quite easy in history_str table as JSON data.
All what it would be necessary to add in zabbix would be only whole apparatus allowing to operate on vector metric and presentation layer extension to present 3D data.