fix(coderd): gate AI task notifications on agent ready state #20690
Conversation
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
Prevents spammy notifications during agent startup by checking that the agent lifecycle state is 'ready' before sending notifications. This ensures notifications are only sent when the agent is fully operational, not during initialization when status may fluctuate. Changes: - Add agent lifecycle check to enqueueAITaskStateNotification - Only send notifications when agent.LifecycleState == Ready - Add debug logging when notifications are skipped - Add test coverage for agent lifecycle states (starting, created, ready) - Update existing tests to explicitly set ready state when expecting notifications Fixes spammy notification reports where users receive alerts before the agent has finished starting up.
johnstcn
approved these changes
Nov 10, 2025
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Relates to coder/internal#1098
Currently AgentAPI waits for only 2 seconds worth of identical terminal screen snapshots before deciding a task has entered a "stable" state. We interpret this as becoming "idle", resulting in a notification being triggered. This behavior is not ideal and is ultimately the root cause of our spammy notifications.
Unfortunately, until we move AgentAPI to either use the Claude Code SDK (or ACP wrapper around it), we are unable to easily fix the root cause.
This PR instead waits until the agent is ready before it will send state change notifications. This will at least resolve some of the complaints about task state notifications being too spammy.
🤖 PR was written by Claude Sonnet 4.5 using Coder Mux and reviewed by a human 👩