Remove UNDEF "Streaming replication connection" Status After Role Change to Primary

[bsislow]

How do we remove the check for "Streaming replication connection" after a cluster has been switched over into a primary?

It shows as the following whether enabled or disabled:

We also removed replication_connection,replication_lag from temboard-agent.conf and restarted the agent, but the status check still shows up.

Thanks.

[bersace]

Hi @bsislow . This is by design. If a check has ever been defined, it is shown as undef once it's disabled. We can't distinguish a check with no data because it is now useless and a check that has no data because of an error or a misconfiguration. We need to review the design for this.

[bsislow]

Thanks, it's just a bit misleading as UNDEF shows up immediately after CRIT and WARN on the overall dashboard and looks suspect even though we can ignore it apparently...

[bersace]

Thanks, that's a good feedback on UX. We'll see what to do on this.

[wahyubudiman]

Hi @bersace , what unit for threshold streaming replication lag base on?

[bersace]

@wahyubudiman sorry, I don't understand your question, can you elaborate ?

[wahyubudiman]

hi @bersace , about the streaming replication lag threshold ,showing 2e+06 for warning threshold, i want to know 2e+06 in byte/minute unit or other unit size? below the screenshot :

[bersace]

hi @bersace , about the streaming replication lag threshold ,showing 2e+06 for warning threshold, i want to know 2e+06 in byte/minute unit or other unit size? below the screenshot :

It's in bytes. temBoard should show the units of this metrics. Marking as bug.

[bersace]

hi @bersace , about the streaming replication lag threshold ,showing 2e+06 for warning threshold, i want to know 2e+06 in byte/minute unit or other unit size? below the screenshot :

Could you please open a new issue about unit in monitoring/alerting ?

[bersace]

Fixed in 8.1