OTA Job Execution Retry Configuration

Hi @Greg_B while I have successfully been using your trick to add the retries to the FreeRTOS OTA jobs, I periodically have the situation that a device hits the maximum of 10 retries. Is there a way to increase that number, or at least to log an event to cloudwatch when this happens? I have been trying to add a Message Routing->Rules with the filter "SELECT * FROM '$aws/events/jobExecution/+'" after advise from Google Gemini but got no events out of it.

Hi. You should get a job execution terminal status message: https://docs.aws.amazon.com/iot/latest/developerguide/events-jobs.html. But you first need to enable JOB_EXECUTION events: https://docs.aws.amazon.com/iot/latest/developerguide/iot-events.html#iot-events-enable. If your job target is a dynamic group, you could use this event to remove the failed device from the group, and then re-add it. This will create a new job execution, with up to 10 new retries.

Are your 10 retries all timeouts, or are there failures? If failures, what's the cause?

Hi Greg, many thanks I missed the part where I have to enable those events, I can see the events now! To answer your question, our OTA is upgrading a Yocto Linux, the download can take a long time and it happens in the background, so for a slow internet connection the user can power cycle multiple times before completion, each time the OTA declares a fail and restarts(resumes) the download. Our job is a dynamic group, the tip of removing and adding again the thing would be gold for us, but I am still failing to identify the number of retries in the failed job execution to isolate the 10th failed retry. I subscribed to $aws/events/jobExecution/# in the console test MQTT but I cannot find executionNumber which I would be looking at (Google Gemini pointed me to it). Other fields I see: {"eventType": "JOB_EXECUTION","operation": "failed","status": "FAILED","statusDetails": {"reason": "rejected: 0x00000010"}}.

The OTA suspend/resume functionality is designed for this purpose, so that you don't get a failure from the interruption: https://www.freertos.org/Documentation/03-Libraries/07-Modular-over-the-air-updates/02-Demos/02-mqtt-ota-agent-orchestrator.

but I cannot find executionNumber

You might have to call DescribeJobExecution: https://docs.aws.amazon.com/iot/latest/apireference/API_DescribeJobExecution.html

Thanks you so much @Greg_B much appreciated