Versions Compared
Key
- This line was added.
- This line was removed.
- Formatting was changed.
Q. How could you restart your bots automatically when they fail?
A.
1) Incorporate a retry step into scenario.
Create a retry step using the Return Code handler and the newly added Failure Handler function.
In the retry step, the Argos API Plugin is placed to start the bot again.
Failures that cannot be detected by the Return Code handler or Failure Handler can be managed with the following plans.
2) Monitor alarms through email and trigger a bot retry.
The Supervisors notification function can send an alarm notification when a bot fails to run.
A bot detecting the alarm starts another target bot.
The target bot can send a success notification to suspend monitoring, when it runs successfully.
This method requires two PAMs.
3) Detect fail or unintended crash by recording the timings of bot start and end in databases or spreadsheets.
This method is effective only if the bot execution time is predictable.
For example, when the bot starts recording, but the record does not appear within 30 minutes, restart the bot.
On-Demand in combination with OpenAPI's PAM Status can avoid collisions.