When I first wrote my automatons, I used sequential but flaky devices would stop the automation. Then I used parallel to avoid stoppages.
The most recent release lets me set continue on error. So if I mark my parallel branch continue on error will any failure of the actions keep the automation from stopping?
Alternatively, should I go back to sequential and mark the actions continue on error. Not really sure on how to approach this.
Thanks
Generally yes.
You’re currently ignoring the problem by just allowing a second to fire. This has always been there btw - it’s just new to the UI. If you don’t NEED parallel I wouldn’t? Fixing your error logic is absolutely the best case.
Go back to sequential for ease of reasoning and predictability, and mark them continue on error.
personally, I’d keep them parallel just because parallel is faster
I would also recommend keeping parallel. I’ve used continue on error for a while, but I’ve had multiple times where it STILL stops on an error (not sure why, but from memory and googling at the time I think it’s sort of up to individual integrations how they handle this?). For automations where I REALLY need it to do all of the actions, I’ve switched them to parallel.
As to the point about fixing the underlying issue, definitely do this if possible. In my case, it’s those very rare and random situations where a device just happens to go unavailable or reports some other error while an automation is running. I’ve had issues specifically with WLED reporting errors even though the device actually did what it was commanded to and never went unavailable.
If the action results in a HomeAssistantError (a specific type in the core python code) then it will continue. Any other type of error it will stop. So yeah, it’s more like “maybe continue if there is an error or not?”