Update configuration of influxdb require a manual restart of site (bug?)

[belese (Blaise)]

Hello,

not a critical bug, but annoying.

When editing influx db connection (move to http for example), in a distributed monitoring. Sites require a manual restart to use new parameters.

Checkmk Enterprise Edition 2.2.0p17
RH8

Kr,

Johan

[Anders]

what do you mean “in distributed monitoring” influxdb is a global setting and the rules as well.

i have never restarted manually, but remember there will not automatically be data to send to influxdb if there are no new metrics, for example if you run the agent ever 5 minute you might have to wait 5 minute until there is data

[belese (Blaise)]

I have a setup with a master and multiple CheckMK slaves (Setup → General → Distributed Monitoring).

Each slave sends its metrics to InfluxDB, as does the master.

If I change the InfluxDB connection on the master and activate it (which applies the change to all sites), it doesn’t work immediately on slave. The slave sites need to be manually restarted for the changes to take effect.

Ps:
If I allow setup on the slaves, I see that WATO on the slaves has the correct values, but in the logs, I see that it still tries to connect to the old ones. Restarting the slave fixes the issue.

[Anders]

Ah!
Now I understand. yea that must be a bug, I’m not 100% sure if it a recent one but we have been sending metrics to influxdb for some time and have quite a lot of slave sites. We have changed the influxdb connectors as well but I cant recall I have seen this issue.

What I can say is that I’m pretty sure the influxdb connection is dropped and started once there are changes that affects the connectors, if that means they still use the old values I’m not sure.