When running a Prometheus Exporter that is connected to a SolrCloud with many nodes/pods, the livenessProbe can fail with a EOF exception no matter what resources the pods are given.
I believe this is due to the kubelet not being able to handle a large amount of data being sent back in a probe response.
In order to fix this we should request back no metrics in the liveness probe. That way the response is empty.
When running a Prometheus Exporter that is connected to a SolrCloud with many nodes/pods, the livenessProbe can fail with a
EOF
exception no matter what resources the pods are given.I believe this is due to the kubelet not being able to handle a large amount of data being sent back in a probe response.
In order to fix this we should request back no metrics in the liveness probe. That way the response is empty.