Failure signature
Azure DevOps-backed tasks fail before tests run with:
TF10216: Azure DevOps services are currently unavailable. Try again later.
The same service outage affected unrelated tasks and PRs within minutes.
Confirmed occurrences
| PR / head |
Build and job |
Failing task |
Outcome |
#26180 / 897e85a4b8e8e206c57ea59a7b0c3088abf18e15 |
14718322 - monotouch_ios, attempt 3 |
Generate BosStorageMirror SAS tokens |
All task retries returned TF10216; the job failed. The stage attempt remains active. |
#26207 / 84ada67506d107a5b8979d2bccb4aa1201b46d3b |
14719170 - introspection, attempt 2 |
Download test libraries |
TF10216 prevented the download; the job failed. The stage attempt remains active. |
#26213 / 6acb80dc4985ae3a81ea8b6487ba2a3ec13148a5 |
14719548 - Publish results, attempt 2 |
Download secrets / OIDC token |
Repeated TF10216 responses failed the publish-results job. The build also had a separate checkout failure after the PR merged. |
First observed: 2026-07-21 23:59 UTC. Latest observed: 2026-07-22 00:04 UTC. Recurrence: 3 jobs across 3 distinct monitored PRs.
Classification and impact
This is an Azure DevOps service/infrastructure outage, not a source or test regression. It blocks credential acquisition, artifact/test-library access, and result publication before the affected work can run.
Investigation
- Inspected each Azure build timeline and failing task.
- Correlated the exact TF10216 response across separate agents, tasks, builds, and PRs.
- Searched open and closed
dotnet/macios issues using TF10216, the full service-unavailable message, and the affected task names; no matching issue was found.
Workaround
Retry failed jobs after Azure DevOps service health recovers, subject to mergeability and retry-budget gates. Do not submit another retry while an equivalent stage attempt is queued or running. Azure logs have limited retention.
Failure signature
Azure DevOps-backed tasks fail before tests run with:
The same service outage affected unrelated tasks and PRs within minutes.
Confirmed occurrences
897e85a4b8e8e206c57ea59a7b0c3088abf18e1584ada67506d107a5b8979d2bccb4aa1201b46d3b6acb80dc4985ae3a81ea8b6487ba2a3ec13148a5First observed: 2026-07-21 23:59 UTC. Latest observed: 2026-07-22 00:04 UTC. Recurrence: 3 jobs across 3 distinct monitored PRs.
Classification and impact
This is an Azure DevOps service/infrastructure outage, not a source or test regression. It blocks credential acquisition, artifact/test-library access, and result publication before the affected work can run.
Investigation
dotnet/maciosissues using TF10216, the full service-unavailable message, and the affected task names; no matching issue was found.Workaround
Retry failed jobs after Azure DevOps service health recovers, subject to mergeability and retry-budget gates. Do not submit another retry while an equivalent stage attempt is queued or running. Azure logs have limited retention.