Trying to troubleshoot what seems to be an Azure service bus issue - the message consumer is logic apps, but this felt like the more appropriate forum.
I completed a load test over the weekend, around 8500 messages processed successfully, but 5 errors. 4 of the errors occurred at very close to the same time, which is why I'm suspicious of some kind of outage.
I have a logic app consuming messages from the service bus - it is configured currently to process up to 10 messages concurrently. For the 5 errors, it appears as if there is some kind of issue when marking the message as 'complete' and hence removing it from the queue - the logic app reports a 'Gateway timeout', and on automatically retrying the same step receives a message from the service bus:
"Failed to complete the message with the lock token '250e397b-f3d1-45e4-9e6f-e07e79ff0037'. The lock on the message has been lost"
On checking the logic app executions, I can see another attempt was made to process the same message with a DIFFERENT message lock ID, approximately 5 seconds after the first attempt encountered the gateway timeout (and hence the 2 logic app executions ran concurrently for a while). It appears that the original lock on the message was somehow abandoned, and the 2nd logic app picked it up from the service bus again.
Summary of some key data for the 2 attempts at processing the message in the image below - the service bus is configured with a lock period of 5 minutes, which I see reflected in the 'enqueued time' and 'locked until' values for the first attempt.
Any ideas? Thanks in advance.