2020-11-28 - aktuelle Serverprobleme

13:20 FIRMWARE UPDATE 1.9.2 for Shelly2.5, Shelly2, Shelly1PM, Shelly1L

This firmware update make devices to report to the cloud in different time which will resolve all delays and cloud disconnects, also they will improve Alexa, Google Home and Smarthings response time.

Do not expect immediately effect on Alexa and Google after update, but control from the APP will be improved right away. To happens at least 30% of devices should be updated to 1.9.2 but same time do not postpone the update, it's a safe.

Another fixes:

Shelly2.5 - Favorite positions could not be stored from local WebUI -fixed

Shelly Flood - Do not execute URL actions in case of Flood - fixed

Thank you Shelly DEV Team for nightwork and fast reaction!

------------------------------------

10:20 Saturday Update:

Tonight we prepare firmware update for Shelly2.5, 1PM, 1L and Shelly2 which resolve the issue. Finally we see that is coming from power measuring statuses which are send at "0" and "30" sec and cause huge load on the servers. Around 14:00 today we will publish it. When most of the costumers update their devices all problems with servers will be gone. !

Update: We found potential problem which can cause devices to disconnecting from the cloud frequently (Reported for many costumers tonight. The same issue cause Alexa and Google slow reaction, time to time.

The problem: It's a little stupid and coming far away but not recognizable so easy because the servers has capacity to hide this problem. Every device pinging the cloud every 30 seconds and tell that is live. But there is a bug and 30 second is starting every time from second "0". Now when we have a thousands device on each server this synchronizing requests from all device start to make a issue.

How we fix that: Tonight we will optimize part of the code on all servers to respond to all request. Tomorrow we will release a firmware fix for at least 2 devices (Shelly1 and Shelly2.5) which will fix the pinging time. Then. each device will report on different time, which is the right way.

Old information:

One of the servers has a issue, we will reboot it. Some of you can expect some short outages.