Hi friends,
Do you have any idea about this? (for example, which metric or expression should I use?)
My dashboard is based on Grafana dashboard #721, with only 3 panels monitoring
- "Messages In Per Topic" sum without(instance)(rate(kafka_server_brokertopicmetrics_messagesin_total{job="kubernetes-service-endpoints"}[5m]))
- "Bytes In Per Topic" sum without(instance)(rate(kafka_server_brokertopicmetrics_bytesin_total{job="kubernetes-service-endpoints"}[5m]))
- "Bytes Out Per Topic" sum without(instance)(rate(kafka_server_brokertopicmetrics_bytesout_total{job="kubernetes-service-endpoints"}[5m]))
--------
I mentioned there are plenty of kafka metrics available but just not sure which should I use.
$ curl -s http://prometheus-prometheus-server.default.svc.cluster.local/api/v1/label/__name__/values |jq . | grep -c kafka_
157
But only 2 for kafka consumers
$ curl -s http://prometheus-prometheus-server.default.svc.cluster.local/api/v1/label/__name__/values |jq . | egrep 'kafka.*cons'
"kafka_server_delayedfetchmetrics_expirespersec_fetchertype_consumer",
"kafka_server_fetcherlagmetrics_consumerlag",
I checked these metrics, but the curves are flat, stuck at 0.
Consumers are far behind the producers, so messages are lost, but I expected to see the lag from "kafka_server_fetcherlagmetrics_consumerlag".
Thanks in advance.