errose28 commented on PR #7552: URL: https://github.com/apache/ozone/pull/7552#issuecomment-2541922103
> If we can identify the performance bottleneck related to the deletion process together, I wonder if we might not need to add too many more metrics. There can always be regressions or unique cases we cannot predict. Just because deletion is working currently doesn't mean we don't need observability into it for the future. > This is because we don’t have a single place that fully describes the DataNode metrics. I've seen a large number of metrics, but I'm not sure what their purpose is. Perhaps we should create an markdown document to specifically explain our monitoring metrics. There's two aspects here: What each individual metric is and how they can be used together to convey useful information For documenting metrics individually, it would be best if every metric had a description attached to it like [this](https://github.com/apache/ozone/blob/e84aa4c4ea7e3d094630bb285afd2f4b38232426/hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/ozoneimpl/AbstractContainerScannerMetrics.java#L39). I'm not sure if hadoop's metric descriptions are preserved in jmx. It would be a good project to try to standardize how we do metrics descriptions and make sure they can be viewed easily from outside the code. We could probably even add a test case that every metric has a description. For demonstrating how different sets of metrics can be used together, I think dashboards provide a sort of self-documenting resource on this. We should have many more dashboards published to go with these metrics. Using these approaches all metrics documentation would stay close to the code, unlike an external document that is more likely to drift from what is actually present. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
