Details
-
Improvement
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
None
-
None
Description
Thank you for adding Prometheus metrics, it's been very helpful in improving the observability of our infrastructure. However we are struggling a bit with them as is due to the way Zookeeper names it's metrics.
The current Prometheus metrics exposed by Zookeeper are not namespaced (ie prefixed) to be zookeeper specific. This makes identifying/exploring the metrics in Prometheus challenging. The Prometheus documentation has the following suggestion regarding metric naming:
...should have a (single-word) application prefix relevant to the domain the metric belongs to. The prefix is sometimes referred to as namespace by client libraries. For metrics specific to an application, the prefix is usually the application name itself. Sometimes, however, metrics are more generic, like standardized metrics exported by client libraries. Examples:
prometheus_notifications_total (specific to the Prometheus server)
process_cpu_seconds_total (exported by many client libraries)
http_request_duration_seconds (for all HTTP requests)
There are a great number of metrics that are specific not to the JVM but to Zookeeper's technology, it would be awesome if those could be prefixed in some capacity. Perhaps simply `zk_`. It would make it more clear which of the metrics are related to Zookeeper the application.
Attachments
Issue Links
- links to