Interrupted workflow sometimes fails due to an unexpected reason

[reubenmiller]

Describe the bug

Flaky test: Resume On Restart A Pending Operation

Resume On Restart A Pending Operation            | FAIL |

'{"@version":"68e3a9d4d81a278de27cfa0c935461558629ec309e010866cb0a567d795e2b8e","duration":120,"logPath":"/var/log/tedge/agent/workflow-sleep-dyn-test-12.log","reason":"sleep killed by signal 15","status":"failed"}' does not contain '"resumed_at"'

https://github.com/reubenmiller/thin-edge.io/blob/main/tests/RobotFramework/tests/tedge_agent/workflows/dynamic_workflow_reloading.robot#L207

Looking at the logs, it seems that it is a legitimate error as the "resumed_at" property is not present in some cases, due to the operation failing due to an unexpected reason, Fail to resume on start: UnknownVersion, and hence missing the "resumed_at" field. Below shows the failed command:

2025-10-20T13:22:10.874509Z      te/device/main///cmd/sleep/dyn-test-12                                 {"@version":"68e3a9d4d81a278de27cfa0c935461558629ec309e010866cb0a567d795e2b8e","duration":120,"logPath":"/var/log/tedge/agent/workflow-sleep-dyn-test-12.log","reason":"Fail to resume on start: UnknownVersion { operation: \"sleep\", version: \"68e3a9d4d81a278de27cfa0c935461558629ec309e010866cb0a567d795e2b8e\" }","status":"failed"}

Failed builds

• https://github.com/thin-edge/thin-edge.io/actions/runs/18652587951?pr=3825
• https://github.com/thin-edge/thin-edge.io/actions/runs/18674195616/job/53241200979