CMK version: CEE 2.1.0p14
OS version: RHEL 7
We’re using the checkmk Grafana Plugin and did create a Dashboard with the important interfaces of our pfSenses.
Now that I see the values in Grafana, I’m not sure if they’re even correct. We’ve chosen bits/sec(IEC) in Grafana. CheckMK uses SNMP but I don’t know how to check what data is actually coming in (bits, or bytes or whatever) - How would I check that?
I pinged Benedikt, the developer responsible for the plugin. Stay tuned for details.
checkmk stores metrics in their canonical form. The interface bandwidth is stored in bytes per seconds. This is then adapted in the graphing system (see cmk/gui/plugins/metrics/network.py and cmk/gui/plugins/metrics/translation.py) but not in the WEB-API which is used by the grafana endpoint to query the data.
If we adapt the WEB-API to return the metric in bits per seconds (as expected) all existing consumer of those endpoints will receive unexpected data (they were used to receiving it in bytes per second).
But we are currently in the progress of writing new API-Endpoints in the REST-API and will consider this input. Perhaps we will add a switch to return the raw data, or the transformed data including units or find another solution.
Unfortunately this will still take some time until it lands in checkmk and the grafana plugin.
Sounds good so far, at least getting the correct information on what data is actually coming in here.
I feel like I didn’t understand correctly though.
CMK internal the bandwidth is stored in bytes but for Grafana via the plugin its bits or do you want it to be bits but its also bytes right now?
My current concept of the situation is the following:
As it is common to express network data in bits/seconds, this situation is not the best, and we should at least add an indicator for the unit in grafana, or better adapt the values on the fly.
Fell free to scatter my concept with some screenshots (values/graphs in grafana vs. in checkmk) of your setup, then we can look into the details.
We discussed this issue internally, and the information I gave you in the previous comments may be wrong.
Could you please show some screenshots of your graphs? We need a value to compare both situations, so reload both grafana and checkmk at the same time and show the latest values:
Also specify which service the values are coming from exactly.
I tried to reproduce this issue with my local installation of Grafana and checkmk and was not able to do so. The values seem to match.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.