You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Local Windows desktop app; one scheduled workflow created and disabled through the app's agent tool surface.
What happened?
Disabling a scheduled workflow does not survive an app restart. Each launch changes its stored enabled value back to true but leaves next_run_at null. The UI therefore shows the workflow as enabled while it never becomes due.
This happened three times in 17 hours on one machine. The stored update followed the app process's first log line by 6 ms on v1.1.10 and 99 ms on v1.1.11:
Stored update
Process start
Delta
2026-08-19T23:09:11.411Z
23:09:11.405 v1.1.10
+6 ms
2026-08-20T19:03:19.124Z
19:03:19.025 v1.1.11
+99 ms
No renderer or CLI session had a transport when the first update occurred: it preceded the process listening on its WebSocket by 816 ms. While the app then remained running for about 20 hours, six workflow-list reads all returned enabled: false with an unchanged updated_at.
The v1.1.10 relaunch used the same version as the preceding process and logged no migration, so this was not introduced by the v1.1.11 schema migration. Only the disabled row's updated_at changed; enabled workflows preserved their timestamps across the same launches.
Enabling through save_workflow normally populates next_run_at; both restart flips left it null. The timing and stored state point to startup handling, but the app source is not public, so that path is inferred rather than code-confirmed.
Steps to reproduce
Create a scheduled workflow with a cron expression.
Disable it through save_workflow.
Quit and relaunch the app.
Observe that it is enabled again but has no next run time.
Expected behavior
The workflow remains disabled across app restarts.
Additional context
Windows 11 build 26310.
The re-enabled workflow never ran. Each occurrence was caught within five minutes against a */30 cron, so none spanned a firing time.
Every disable in this sample used save_workflow; disabling through the Workflows UI was not tested.
Only one disabled workflow was observed across restart.
[Filed by Copilot on behalf of @bghgary]
Affected version or release
v1.1.10 (
a7315e1) and v1.1.11 (0155271)Installation context
Local Windows desktop app; one scheduled workflow created and disabled through the app's agent tool surface.
What happened?
Disabling a scheduled workflow does not survive an app restart. Each launch changes its stored
enabledvalue back to true but leavesnext_run_atnull. The UI therefore shows the workflow as enabled while it never becomes due.This happened three times in 17 hours on one machine. The stored update followed the app process's first log line by 6 ms on v1.1.10 and 99 ms on v1.1.11:
2026-08-19T23:09:11.411Z23:09:11.405v1.1.102026-08-20T19:03:19.124Z19:03:19.025v1.1.11No renderer or CLI session had a transport when the first update occurred: it preceded the process listening on its WebSocket by 816 ms. While the app then remained running for about 20 hours, six workflow-list reads all returned
enabled: falsewith an unchangedupdated_at.The v1.1.10 relaunch used the same version as the preceding process and logged no migration, so this was not introduced by the v1.1.11 schema migration. Only the disabled row's
updated_atchanged; enabled workflows preserved their timestamps across the same launches.Enabling through
save_workflownormally populatesnext_run_at; both restart flips left it null. The timing and stored state point to startup handling, but the app source is not public, so that path is inferred rather than code-confirmed.Steps to reproduce
save_workflow.Expected behavior
The workflow remains disabled across app restarts.
Additional context
*/30cron, so none spanned a firing time.save_workflow; disabling through the Workflows UI was not tested.