context canceled error when collecting metrics from thanos-querier

[LeonD9]

Expected Behavior

No errors in log of kube-metrics-adapter

Actual Behavior

I receive the following error after the collector runner is stopped for connecting to the thanos-querier:

time="2024-08-15T11:20:33Z" level=info msg="stopping collector runner..."
time="2024-08-15T11:20:33Z" level=error msg="Failed to collect metrics: Post \"http://thanos-querier.kube-system.svc.cluster.local:10902/api/v1/query\": context canceled" provider=hpa

Is this ERROR expected? it happens frequently but i didn't see any impact on HPA, what can i do to resolve it?

Steps to Reproduce the Problem

Upgrade from v0.1.8 to v0.2.3, i didn't see this kind of error messages in v0.1.8 but i started seeing them after i upgraded to v0.2.3
Run kube-metrics-adapter with --prometheus-server=http://thanos-querier.kube-system.svc.cluster.local:10902 arg

Specifications

Latest image version: ghcr.io/zalando-incubator/kube-metrics-adapter:v0.2.3
EKS: v1.24

[mikkeloscar]

This has likely been introduced with: #713

The collectors are stopped when the HPA is changing and the context is cancelled.

This should not have an impact on the functionality, but I agree it's a bit confusing to have such logs.