We are providing updates here:
Yesterday (07/24), my Smart Lighting “at sunset -20 min” rule fired at 5:32 PM, about 3 hours ahead of schedule. If this isn’t a scheduler problem, then what?
commandId 288
date 2016-07-24 5:32:00.138 PM PDT (2016-07-25T00:32:00.138Z)
description Living Room Lighting Automation sent on command to Living Room Sconce
deviceId 7e950796-3e8e-4e2e-906b-89a76c44f5e9
displayed true
eventSource APP_COMMAND
Putting a thought out there, is ST using Amazon Lambda? Lamda is supposed to a event driven platform that automatically scales and high reliability etc. Wonder what Robert thinks about it
Have the “weekly updates” stopped? Or did I miss them?
FWIW, for a long time I had no scheduling issues. Recently, a number of Smart Lighting automations began failing. Support says they are well aware of the problem & engineering is looking into it – but an occasional update from the top might help keep current customers in the fold…
Or maybe he’s about to give up on the platform like the rest of us.
About to? Now that’s funny!
I can confirm this too. Some Smart Lighting automations don’t fire. Occasionally a routine won’t fire. Usually my goodnight and/or evening routines
Same here. Have noticed 3 routines in the last couple days that did not fire at the scheduled time. They’ve been solid for months. Two were sunset/sunrise based, one was time-based.
We’re seeing some server hotspots, we are looking into the cause. I hope to have some updates soon.
Mine are time based and one is rule is contact sensor triggered… Lights come on when door is opened, turn off when door closes, after 3 minutes.
Thanks for the reply! Hope it’s sorted soon
We saw a increase in scheduling activity, we are expanding the cluster in the short term until we can find the root cause.
If I am not mistaken expanding the cluster was the fix the last time around, or was it the first time?. It always seemed to me that timed event failures was the direct result of capacity issues.
It’s a little bit different this time. But yes, last time was capacity issues, the increase here we think is reversible.
It does appear that the rate of timed event failures have diminished over the past 24 hours.
Both Alex and I were out last week. We mentioned in a post 2 weeks ago that they wold resume when he was back.
Maybe I missed it he still on vacation 2 weeks later?