I have one from April 3, another from April 10 (which I looked into a bit), more from July 14, 16, and 17, some from Oct 7 and 8, and now one from today.
You are asking why your Post-Login Action periodically fails with an actions_execution_failed event and an error_class of platform_error, and why the log entry gives no detail about the cause.
Root Cause: A platform_error usually means the Action stopped before it finished. The most common cause is an unhandled exception or a rejected Promise inside the Action, such as a failed call to the Management API or an external service that the Action does not catch. The log entry records only that the execution failed, so the specific error must be found in the Action's own logs.
Solution: Make every asynchronous call in the Action handle its own errors, then find the detailed error for each failed execution.
Locate the failing execution in the Dashboard logs. Use the execution_id (68c72e52-44d3-4320-b66d-cc192f6937a0 in the example you shared) and the action_id from the event to find the matching Action run. Open the Action's real-time logs or the Action execution logs for the same time window.
Addawait to every Promise-based call in the Action, including Management API calls and HTTP requests. A call without await can fail after the Action has already returned, and the failure is reported as a platform error.
Wrap each external call in a try/catch block. In the catch block, call console.error with the error message so the failure appears in the Action logs, and decide whether to allow the login or deny it with api.access.deny.
Set a timeout for external calls. If your backend or a third-party service is slow, the Action can exceed its execution limit and fail. Keep each call short and handle the timeout explicitly.
Confirm the Action's secrets and dependencies are set and that the Action version in the event (59ca304f-4873-4a9d-8fe5-0799ad8c04b3) matches the version you deployed.
Please follow up if you need further assistance reviewing the Action code.