fix(deploy): stop the reconcile job uninstalling the chat - #36
Merged
Conversation
The job uninstalls any `mcp-*` release with no `toolsets/<name>` directory behind it, skipping `mcp-index` — but not `mcp-chat`, which is not a toolset either. So every deploy ran `helm uninstall mcp-chat`, and the chat only came back because the `chat` job reinstalls it a minute or so later. Both jobs only need `detect`, so they race, and reconcile has no image to build and always wins. In the 8 September deploy the chat was uninstalled at 14:30:29 and reinstalled at 14:32:05 — 95 seconds down, conversations lost, on every deploy. A failed or cancelled `chat` job would have left it down. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The reconcile job uninstalls any
mcp-*release with notoolsets/<name>directory behind it. It skips
mcp-index, but notmcp-chat, which is not atoolset either — so every deploy uninstalled the chat.
It came back only because the
chatjob reinstalls it. Both jobs need nothingbut
detect, so they race, and reconcile has no image to build and always wins.Evidence from the last two deploys, both logging
Uninstalling mcp-chat: toolsets/chat no longer exists, thenrelease "mcp-chat" uninstalled:So
chat.<host>was down for about 95 seconds on each deploy, losing everyin-memory conversation, and a
chatjob that failed or was cancelled wouldhave left it down until someone noticed.
Found while confirming that removing a toolset directory removes its service.
It does — this is the same mechanism removing something it shouldn't.
🤖 Generated with Claude Code