Matt Burgess created NIFI-7273:
----------------------------------
Summary: Add flow metrics REST endpoint with for Prometheus
scraping
Key: NIFI-7273
URL: https://issues.apache.org/jira/browse/NIFI-7273
Project: Apache NiFi
Issue Type: New Feature
Components: Extensions
Reporter: Matt Burgess
NiFi has the ability to expose endpoints for Prometheus to scrape via the
PrometheusReportingTask (NIFI-4362) and via components that use the
PrometheusRecordSink controller service. However that involves adding
components to the overall flow, which requires their own configuration and ends
up generating their own metrics that contribute to rollup metrics and queries.
This Jira proposes to add an endpoint to the NiFi REST API that exposes the
following metrics/information in Prometheus format for scraping:
- Root Process Group status (recursive to include all components)
- Connection Status Analytics (backpressure predictions, e.g.)
- JVM Metrics
- Bulletins (for use by AlertManager, not necessarily a metric per se)
Standard security/authorization principles apply, and it is proposed to offer
node-specific metrics rather than cluster-wide aggregates, as Prometheus can
then choose how to do the aggregates as necessary.
It may be prudent to refactor PrometheusMetricsUtil out into its own module,
for use by the various components in various modules (to now include the
framework).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
