Open nicolasparada opened 1 year ago
I'd be inclined to say it should fail with a config error here: you may be relying on that alias for alerting or other monitoring so silently altering it (e.g. to another ID) is not a good idea as now you will be watching the wrong metrics.
This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 5 days. Maintainers can add the exempt-stale
label.
This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 5 days. Maintainers can add the exempt-stale
label.
This still happens even on v2.0.10.
alias + instance id simliar to other plugins is my vote
I'm not a fan of instance ID as it is not consistent though, particularly when you use @include
with wildcards which deliberately does not enforce any ordering. It means you can run the same config twice (particularly from configmaps in K8S) and get two different orders for instance IDs.
Using the same alias should be a failure I think.
Bug Report
Describe the bug
When using the same alias for the same plugin instance, only metrics for the last matching instance are produced.
To Reproduce
Given this config:
I see metrics
/api/v1/metrics
with only onesame
plugin with10
records per second. Then if I change the order:Now I see metrics with only one
same
plugin again, but with20
records per second.Expected behavior
Maybe it should not allow giving the same alias. Or it should use the internal instance ID to report metrics when there are more than one instance with the same alias.
Your Environment
cr.fluentbit.io/fluent/fluent-bit:latest