Since last week (this is quite recent), I’ve seen the same exception in the logs again and again for my Custom Ecobee Device:
Java.util.concurrent.TimeoutException: Execution timed out after 40 units. Start time: Tue Jun 16 22:20:15 UTC 2015 @ line 1483
And, the following exception, for some smartapps,
Java.util.concurrent.TimeoutException: Execution timed out after 20 units.
This is a new issue that I’ve not seen before.
I think that the rate limiting is becoming too restrictive. Can somebody at SmartThings look into it? I’ve opened a support ticket (#106962), but support is pretty overwhelmed right now (based on some other tickets I have).
Does your ecobee device have something that runs for a long time? I’ve only seen the timeouts happen when there is something in the code that is running unusually slowly.
Disclaimer: I don’t have much actual knowledge of how we handle rate limiting and timeouts
Now, it seems that the rate limiting for device handlers is down to 20 units from 40.
As 20 units = 20 seconds, for cloud-to-cloud integration such as MyEcobee and MyAutomatic device handlers, it is starting to be really limiting , and there are more exception in the logs such as
Java.util.concurrent.TimeoutException: Execution timed out after 20 units.
I think that there should be a different rate for those cloud-to-cloud integrations.
Hi @tslagle13, I’m seeing more and more exceptions regarding rate limiting in my logs.
Will ST reconsider the rate limiting for cloud-to-cloud integrations (such as Ecobee, Automatic, Neurio, etc.) which typically take longer than others to execute?
Those integrations are dependent on the partner’ latency, the network latency,and there may be a lot of factors which impact the delays involved (connection issues, cloud infrastructure, APIs design, etc.)
I think that there should be some specific rate limiting thresholds for cloud-to-cloud integrations that are different from the rest.