Enhance the newly released “runtimeMetadata” object sent with integration actions to include the number of retries remaining before monday considers it a failure.
This way we know when we should return a meaningful error message (if possible) rather than allowing another retry. (For example a third party system is having connectivity issues, we might allow monday to retry but eventually need to report back to the user “external system is unresponsive”).
We can count the times we’ve seen it, but its entirely possible due to an error we won’t see every time to count correctly - so that is not the most reliable way to count. It would be much better for monday to send us a definitive count or remaining retries so we can behave appropriately.