Skip to main content
RisingWave exposes Prometheus metrics from each component, including compute, frontend, compactor, and meta nodes. If you need the complete metric catalog instead of the curated Grafana dashboards, use the source-maintained reference below.

​
Metrics reference

Use the RisingWave metrics reference for the complete Prometheus metric documentation. It covers:
  • Default metrics endpoints and ports
  • Common labels and join/info metrics
  • Alert thresholds used by the shipped Grafana dashboards
  • Operational metrics grouped by domain, such as streaming, storage, sources, sinks, memory, CPU, batch queries, and UDFs
For table change log troubleshooting, the reference also includes these metrics:
  • state_store_table_change_log_fetch_latency: latency for fetching table change logs from the state store.
  • state_store_table_change_log_cache_counts{result="hit|miss"}: cache hit and miss counts for table change log lookups.
  • storage_table_change_log_get_latency: latency for serving table change log requests on the metadata path.

​
Metrics inventory

Use the generated metrics inventory TSV when you need the flat list of registered metric families. The inventory is extracted from the RisingWave source code and includes these columns:
  • metric_name
  • type
  • labels
  • help
  • buckets
  • file
The reference and inventory are maintained in the risingwave repository, where CI checks keep the inventory aligned with the current metric registrations.