Zabbix 4.2.8 - Two Graphs for Same Metric Showing Differing Information

[alexw-z]

Hi

I am looking after a Zabbix 4.2.8 server, and an unusual issue has been reported with a discrepancy in the data showing in two different graphs.

The metric itself is fairly custom but simple, and displays a count of TCP connections on a certain port once every 30 seconds.

system.run["ss -t | grep '172.31.8.140:11211' | wc -l"]

One graph is a widget on a Dashboard:

https://i.imgur.com/1GI57BM.png

the other graph is situated within Monitoring < Graphs:

https://i.imgur.com/V5f1dod.png

(note I am referring to the data, not the different in screen size)

I’ve made copies of both and played around with the settings, but mostly of these seem to relate to display options and I’ve not found anything that makes any dramatic changes in the content of either.

I am aware the version of Zabbix is out of date, and I am working on an upgrade to bring the server up to 5.0 in the coming weeks.

In the interim however, I’ve done some googling for potential known bugs, but I’ve not found anything definite so far.

Is anybody aware of any (presumably historic issues of this type), or can anybody offer any ideas if there’s anything else I can do to try and resolve this problem?

Thanks.

[[email protected]]

Looking at these graphs, it would seem to me like you're trying to define the Y scale as being 35 - 193 vs .2 - 1.6k
The first scale is going to show every small fluctuation, and the second scale is going to smooth that out.

What happens when you change the graphs to have the same Y scale?