Too many deploys
Injected failure: per-agent hourly total exceeded. Three deploys to other services this hour, then a fourth: each is fine alone, four in an hour is not.
Mode too_many_deploys of the failure matrix, on one of its eight workflows. Each run below is the sandbox's scripted agent acting on the sandbox's fake systems of record under the injected failure, an oracle reading those systems' databases, and the verdict of engine 1.0.0 from the systems of record alone.
Last changed Markdown versionllms.txt
| Workflow | Agent reported | Oracle | CommitLayer | Why |
|---|---|---|---|---|
| Deploy a service | success | FAIL | FAIL | This agent has deployed more than three times in an hour (was 4) |
Deploy a service
Scenario deploy/too_many_deploys: per-agent hourly total exceeded.
Earlier actions by the same agent, run first: svc_a (version=1.0.1, environment=production); svc_b (version=1.0.1, environment=production); svc_c (version=1.0.1, environment=production).
What the agent reported
The agent reported success. Its calls, and what it saw come back:
- GET /platform/services/svc_api200
- POST /platform/services/svc_api/deploy200
- POST /platform/services/svc_api/healthcheck200
- POST /crm/notes200
Before these, the earlier actions made 12 calls; each came back 200.
What the systems of record showed
Records written after the action:
- platform.deploy_created on svc_apiversion=1.5.0, environment=production
- platform.health_check on svc_apistatus=healthy, version=1.5.0
- crm.note_added on svc_apiauthor=lab-agent, text=Deployed 1.5.0 to production
The oracle, reading the databases directly: FAIL
aggregate: more than 3 deploys in an hour
The verdict and why
FAIL This agent has deployed more than three times in an hour (was 4)
Outcome PASS, policy FAIL; the oracle reads outcome PASS, policy FAIL.
- aggregateThis agent has deployed more than three times in an hourcount of platform.deploy_created over 1h per agent above 3
Run it in the Failure Lab
The Failure Lab replays each run above: the agent's calls, the records the systems held and the verdict. No account is needed. The verdict it shows is the one recorded with the run: the verdict counts the agent's earlier actions, and the dry run the lab sends verifies one action.
The sandbox in the app runs the exact scenario in your own workspace, as a verified action: