markhoerth opened a new pull request, #12402: URL: https://github.com/apache/gravitino/pull/12402
HealthOperations and HealthAliasServlet ship in the server but have no doc page, so the endpoints, the root aliases, and gravitino.server.health.entityStore.probeTimeoutMs are undiscoverable outside the source. Covers what readiness actually tests, an existence lookup for a sentinel metalake run on a bounded thread pool rather than on the request thread, and the five reasons a check reports DOWN, including probe-rejected, which indicates probe traffic outpacing the store rather than any single probe failing. Also notes that the Helm chart's default probes target / rather than the health endpoints, so a pod whose entity store has failed still reports ready. Correcting the chart defaults is left to a separate change, since the health config entry is marked 1.3.0 and pointing chart defaults at it would need a version floor in Chart.yaml. Every claim is checked against source: the three endpoints and status codes, the four root aliases with /health.html mapping to the aggregate, the httpServer and entityStore check names, the gravitino_health_probe sentinel, the 1/4/20 probe thread pool, the five DOWN reasons, and probeTimeoutMs defaulting to 2000. -- 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]
