CI flake: unity-testsuite

[clockwork-labs-bot]

Detected by CI Gremlin after a failed GitHub Actions attempt succeeded on rerun.

Tests:

• unity-testsuite

Error:
The job has exceeded the maximum execution time of 30m0s

Failed job runs:

• https://github.com/clockworklabs/SpacetimeDB/actions/runs/29064277650/job/86274837342 (unity-testsuite, attempt 1 -> 2)

Timeout detection:

• Source: check-run annotation
• Inferred: no
• Detail: The job has exceeded the maximum execution time of 30m0s

Context:

• none

[clockwork-labs-bot]

Diagnosis for the linked job attempt:

I inspected only the exact linked failed job: run 29064277650, job 86274837342, unity-testsuite, attempt 1.

The job did not fail on a Unity assertion or a SpacetimeDB test failure. It was cancelled by the job-level 30 minute timeout:

The job has exceeded the maximum execution time of 30m0s

Relevant timeline from the exact job:

• Job started at 2026-07-10T02:38:36Z on ghr-github-self-hosted-ubuntu-14900k-01-spacetimedb-14-d955e430.
• Setup consumed most of the budget before Unity tests started. Notably, checkout took about 84s, Check v8 outputs took about 12m15s, and installing the local CLI took about 3m51s.
• Start SpacetimeDB and Publish unity-tests module to SpacetimeDB both succeeded by 02:58:35Z.
• The Unity Library cache missed:

Cache not found for input keys: Unity-shub/fix-replay, Unity-

• Run Unity tests started at 02:58:35Z, pulled unityci/editor:ubuntu-2022.3.32f1-linux-il2cpp-3, activated Unity successfully, then entered a cold Unity project import/rebuild:

Rebuilding Library because the asset database could not be found!
Packages were changed.
The following packages were added:
  com.unity.ai.navigation@1.1.5
[Package Manager] Done resolving packages in 134.52 seconds
...
Begin MonoManager ReloadAssembly
- Loaded All Assemblies, in  0.185 seconds
Native extension for LinuxStandalone target not found

• The next event was cancellation at 03:08:36Z:

##[error]The operation was canceled.

Root cause for this exact attempt: the Unity job hit the 30 minute workflow timeout while the Unity test runner was still doing cold editor/package/import startup work after a Unity Library cache miss. The log does not show a playmode test failure, Unity licensing failure, Docker pull failure, or SpacetimeDB server failure.

Suggested title:

CI flake: unity-testsuite times out during cold Unity Library rebuild

Duplicate check:

Similar but not close enough to mark duplicate:

• CI flake: unity-testsuite - access token is unavailable #5302 / CI flake: unity-testsuite #5304 / CI flake: unity-testsuite #5408 / CI flake: unity-testsuite #5409: Unity licensing/access-token flakes.
• CI flake: unity-testsuite #5281 / CI flake: unity-testsuite #5268: Unity authentication flakes.
• CI flake: unity-testsuite - 502 pulling Unity editor #5282: Docker Hub 502 while pulling the Unity editor image.
• Run Unity tests under cargo test #4924: broad request to run Unity tests under cargo test, but it is not an exact timeout/import flake.

Top duplicate candidate: none. I did not find an existing Test Flakes ticket for unity-testsuite timing out during cold Unity Library rebuild/import.