SmartThings Outage - Mar 12 2018

[WB70]

@Zach_Varberg

I mostly just chimed in on this thread because I wanted to correct some misinformation about what scheduling functionality runs on the hub.

The messenger always gets shot at

@Zach_Varberg

And I DO want everything to run locally, it can just be frustrating as a developer to be working on improving something and to still get nothing but complaints about it

The majority of the community appreciates the work you guys do and know that a lot of you share the same sentiments on things that might be lacking from a customer perspective.

[joecooler]

@WB70

@Zach_Varberg

I mostly just chimed in on this thread because I wanted to correct some misinformation about what scheduling functionality runs on the hub.

The messenger always gets shot at

@Zach_Varberg

And I DO want everything to run locally, it can just be frustrating as a developer to be working on improving something and to still get nothing but complaints about it

The majority of the community appreciates the work you guys do and know that a lot of you share the same sentiments on things that might be lacking from a customer perspective.

What he said.

[ogiewon]

@jsulliweb

Gonna ask a question that you may not be able to answer (and may have been asked before, the thread is kinda long at this point), but I’m curious. I had several webcore pistons that just handle basic lighting and such that still worked during the outage. That’s the first time that’s ever happened. Is there a reason that they worked this time?

If you have the time to read through this whole thread, there are some clues as to why automations, even cloud automations, continued to work during the outage. It seems the major cloud service that was overwhelmed was the Authentication Services. Thus, no new authentications could be completed (or very few) and users could no longer log into the ST Mobile App or ST Web IDE. However, SmartApps that were already running (e.g. webCoRE) were able to continue to run and interact with the end-user’s hub.

Here are the posts from earlier in the thread that provides the clues.

SmartThings Outage - Mar 12 2018

Our monitors on subscriptions and schedules are looking fine. The mobile clients, IDE, and SmartApps that rely on API calls are the things primary affected by this outage.

SmartThings Outage - Mar 12 2018

I sent you a DM. I would have expected that Smart Lighting, along with CoRE/webCoRE should have still worked in most use cases based on some of our monitors (there might be some edge cases).

[tgauchat (ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)]

@ogiewon

If you have the time to read through this whole thread, there are some clues as to why automations, even cloud automations, continued to work during the outage…

You’d think that SmartThings would make an effort to publicize exactly this fact. It’s only a bit of a consolation - but a genuine one. All sensor or schedule triggered lighting, for example, continued to operate properly both local and cloud instances.

[Lighty]

I do understand you all have your issues with the cloud…

Not cool with these latest problems, but I am still concerned with the presence sensors, they still do not work after 5 days now…

I am one of those who are considering another alternative product, I will not name, due to it being to new.

[Lgkahn (Larry)]

For routines with presence sensors I have found that removing and redoing the routines seems to fix it.

[gausnes (Luke - Backend Engineer)]

I believe there should be a more official statement coming, but we try to isolate critical functionality the best that we can to minimize single points of failure. I sent the following in a DM to someone that was curious during the outage.

WebCore/Smart Lighting/CoRE execute in clusters that aren’t dependent on API interaction. Executions for these SmartApps generally use internal schedules or device events to prompt executions. This type of activity uses different forms of authentication than API interaction does. This enables us to break these nodes apart into different clusters since they have different dependency on authentication along with other differing aspects of those types of execution. This allowed the clusters that perform SmartApp executions triggered by subscription and schedules to not as affected by the API auth service that went down earlier today.

[coolVariable]

This still has not been resolved for me.

And ST support is doing NOTHING. WTF!

[coolVariable]

HUBITAT?

Seems like that is the one that everyone is recommending due to ST having turned into a MESS.

[gausnes (Luke - Backend Engineer)]

What are you seeing? The outage that this thread is about has be resolved for awhile.

[jkp]

Ha ha good luck with getting a response from @coolVariable

No working devices since Friday 3/16 - NO RESPONSE from Support

This is getting beyond ridiculous. Every couple of days NOTHING works. Since Friday 3/16 NOTHING WORKS and there has been NO RESPONSE from support. Is this a JOKE? About to call the credit card company to make a claim against Samsung SmartThings.

[Mindstates (Josh)]

This still has not been resolved for me? Any status update?

[kahilzinger (Kirk Hilzinger)]

Call/E-Mail support if powering off the hub with the batteries out for 15 minutes does not work.