Stabilize sandbox Elasticsearch StatefulSet#1013
Open
Conversation
added 3 commits
March 19, 2026 18:02
- Remove opensearch-deployment.yaml and opensearch-pvc.yaml (unused, no app references) - Set elasticsearch replicas to 1 and configure discovery.type=single-node
Missing proxy-body-size annotation caused nginx to use its default 1m limit, rejecting large publish payloads (~3.4MB) with 413 errors. Matches the limit already set in sandbox and staging.
- Set podManagementPolicy to Parallel so nodes can start simultaneously and discover each other during cluster formation. OrderedReady caused a deadlock where neither node could become ready without the other. - Set timeoutSeconds: 10 on readiness and liveness probes. The default of 1s was too short for the wait_for_status=yellow health endpoint, causing nodes to be killed before they could join the cluster. The elasticsearch-discovery service already had publishNotReadyAddresses: true in the manifest but was not applied to the live cluster (drift). This has been applied directly.
edgarf
approved these changes
Mar 26, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
podManagementPolicy: Parallel— nodes start simultaneously and can discover each other during cluster formation.OrderedReady(the default) caused a deadlock where neither node could become ready without the other already being up.timeoutSeconds: 10on readiness and liveness probes — the default 1s was too short for thewait_for_status=yellowhealth endpoint, causing nodes to be killed before they could join the cluster.The
elasticsearch-discoveryservice already hadpublishNotReadyAddresses: truein the manifest but had drifted from the live cluster. This has been applied directly.These fixes were identified during an incident where a dead EC2 node (kubelet stopped posting status) caused the cluster to lose quorum and enter a CrashLoopBackOff cycle that could not self-recover.