FAQ: Second execution for throttled invocations doesn't have the correct delay

Hi. I counted for six minutes delay when setting it up in throttle parameters. And it works correctly for all executions except the second one. Could you please advise, how can I fix it?


@hexlet , good catch, this behavior is expected given the implementation of our throttling logic (I gave this feedback to the team so we’re aware and can improve this in the future).

You can find more information and one potential solution here.

This solution worked, thank you.

But I’ve had to separate reading and saving timestamp value to different steps. For unknown reasons $checkpoint contained undefined after all steps executed, and I couldn’t use it in the next workflow. Why would this happen?

I found out that it caused by “GET request” action. Replacing it to “HTTP/Webhook request” fix the issue