[HDDS-1811] Prometheus metrics are broken for datanodes due to an invalid metric

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 0.5.0
• Component/s: Ozone Datanode
• Labels: pull-request-available
• Target Version/s: 0.4.1
• Sprint: Done
• Assignee: Attila Doroszlai
• Reporter: Marton Elek
• Created: 16/Jul/19 16:32
• Updated: 24/Feb/20 20:53
• Resolved: 22/Jul/19 21:35

Description

Datanodes can't be monitored with prometheus any more:

level=warn ts=2019-07-16T16:29:55.876Z caller=scrape.go:937 component="scrape manager" scrape_pool=pods target=http://192.168.69.76:9882/prom msg="append failed" err="invalid metric type \"apache.hadoop.ozone.container.common.transport.server.ratis._csm_metrics_delete_container_avg_time gauge\""

Issue Links

• incorporates: HDDS-1835 Improve metric name for CSM Metrics
• is related to: HADOOP-16398 Exports Hadoop metrics to Prometheus
• relates to: HDDS-1867 Invalid Prometheus metric name from JvmMetrics
• links to: GitHub Pull Request #1118