fix(copilot): Use state field instead of status for task lifecycle

[JoshFerge]

The GitHub Copilot Tasks API returns task lifecycle values (queued, in_progress, completed, failed, timed_out) in the state field, while the status field is always an empty string. Our polling code was checking task_status.status == "completed" which was never true, so Copilot tasks would stay stuck in RUNNING state forever after completion — the PR URL and completion state were never synced back to Sentry.

Validated by hitting the live Copilot API: created a task, polled until completion, and confirmed that state carries the lifecycle values while status is always "". The partner API spec we were given listed status as the field, but the actual API uses state.

Changes:

• Add state field to GithubCopilotTask model (keep status for forward-compat)
• Switch polling logic to check task_status.state instead of task_status.status
• Update test mocks to match real API response shapes (state, in_progress, github_resource, updated_at)

Agent transcript: https://claudescope.sentry.dev/share/7RfoUCo3OmihUbUrznGRpJ0bE9LXTfcbFamSXhLirtc

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Model field name doesn't match real API response
  • Renamed the model field from last_updated_at to updated_at to match the real GitHub Copilot API response shape that the test mock uses.

Or push these changes by commenting:

@cursor push 3f7e5a5018

Preview (3f7e5a5018)

diff --git a/src/sentry/integrations/github_copilot/models.py b/src/sentry/integrations/github_copilot/models.py
--- a/src/sentry/integrations/github_copilot/models.py
+++ b/src/sentry/integrations/github_copilot/models.py
@@ -77,7 +77,7 @@
     session_count: int | None = None
     artifacts: list[GithubCopilotArtifact] | None = None
     archived_at: str | None = None
-    last_updated_at: str | None = None
+    updated_at: str | None = None
     created_at: str | None = None
     sessions: list[GithubCopilotSession] | None = None

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}

[cursor]

Model field name doesn't match real API response

Low Severity

The test mock was updated from "last_updated_at" to "updated_at" to match the real Copilot API response shape, but the GithubCopilotTask model field is still named last_updated_at. Since Pydantic ignores unknown fields by default, the API's updated_at value is silently dropped and last_updated_at will always be None when parsing real API responses. This is the same class of bug this PR fixes for status/state — just for a different field.

Additional Locations (1)

• tests/sentry/integrations/github_copilot/test_client.py#L55-L56