search

signalfx / signalfx-k8s-metrics-adapter

Horizontal Pod Autoscaler custom/external metrics provider for Kubernetes that uses SignalFx as the backend
Apache License 2.0
9 stars 15 forks source link

[NoIssue] Question about errored_jobs #14

Closed rjmsilveira closed 3 years ago

rjmsilveira commented 3 years ago

Hi, this is not necessarily an issue, but more a question.

Was testing the metrics exposed and noticed that I cannot make the metric errored_jobs be triggered.

I tried to make syntax errors within signalflow, setting all signals to false, and even make the Adapter fail to connect to SFx but I don't see it being activated.

Can you help me understand when should it be triggered?

Thanks in advance

keitwb commented 3 years ago

Are you collecting the metrics with the Smart Agent? The adapter doesn't actually send metrics about itself directly, it just exposes them.

Also check if the logs have a message like SignalFlow job errored, restarting in a bit:. If this displays, then there should definitely be some errors registering for the errored_jobs metric. Also note that that metric is a cumulative metric so by default in charts it will show as a delta, which means it only shows the diff from the last interval sent.

rjmsilveira commented 3 years ago

Yes, we use Smart Agent polling metrics from adapter. Meanwhile errors happened and we got alerts as expected. Thanks