Failed Routines - possible explanation revealed by live log

[geko]

This issue has been there forever. Search for TimeoutException and you’ll find a dozen of discussions. For example, this:

Unschedule() API Execution Time Too Long?

This is a question for SmartThings software developers. During smart app development I’ve run into a problem when one of my functions occasionally triggered java.util.concurrent.TimeoutException which means that it’s exceeded 20 sec execution limit. With some debugging I was able to narrow it down to calling the unschedule() API. A simple test revealed that executing unschedule() API can take anywhere from 90 milliseconds to 3.7 seconds (!). And that’s on a good day. Occasionally, unschedule() gets stuck for longer than 20 seconds, resulting in java.util.concurrent.TimeoutException. Here’s a simple test that I ran: private def myUnschedule() { def t0 = now() unschedule() def t = now() - t0 log.trace "unschedule() is executed in ${t} ms" } And here’s what it prints: X…