Loss of State affecting some users / some SmartApps (was Death knell for RM today?)

Coincidence, Nah, I don’t think so. Let’s go through some facts:

  • My Netatmo integration with other platforms (VeraPlus, ImperiHome…) did not have any issues and is still working, and they have not released any updates, so it looks like the Netamo API hasn’t changed and it’s working;
  • ST Support told me that some users were able to login/connect again using the Netatmo (connect) smart app and that this resolved the issue, and also since this hasn’t affected all ST users that use Netatmo it looks like the Smart App and Child DH codes are still working;
  • I’ve reported to ST Support that I’m unable to login/connect to Netatmo because when I try it’s returning and XML page with the message “error=invalid_token”;
  • Curiously there is a variable called “accessToken” and it’s stored on the smart app Application State.
    Let’s see: does it look like there was an issues with the Application State of this Smart App? Just saying…
6 Likes

Déjà vu circa April…

History repeats itself again :unamused:

5 Likes

Have I missed it?

7 Likes

@a4refillpad

You missed that everything from ST, while may seem specific to a reasonable person, is actually relative.

For example, the term Weekly.

This is culture. Adopted by the C suite.

4 Likes

Hey, bi-weekly has the word weekly in it, which is better than biannual, which only recently replaced annual. But the reality is, that there shouldn’t be any updates if things would roll like they are supposed to. We don’t need updates on f-ups, we just need no f-ups.

5 Likes

It’s coming. We’re ironing out some details for the full post mortem.

2 Likes

known bug, will get fixed. Eventually. It’s in the javascript realm…

“When you eliminate the impossible, whatever remains, however improbable, must be the truth.” - Leonard Nimoy

2 Likes

Actually Spock quoting Sherlock Holmes. Credit where credit is due and all that. :wink:

7 Likes

Haha, leave it to JD to dig up the facts! You are the real authority :wink: glad you are here!

5 Likes

Please let us know when we can expect SmartThings to stay stable for at least one month. Based on history, that has never happened.

http://status.smartthings.com/history

Thanks for your time.

I don’t understand the question…? Don’t you think SmartThings is trying to avoid outages? Do you think they have a crystal ball to know exactly when a specific target level of reliability will be reached? :crystal_ball:

Complex systems inherently have reliability issues. Except for “anomalies” or specific explained reasons for steps-backwards, the best way to predict when a particular level of reliability will be reached is by extrapolating from a trend-line of what should be a steadily increasing “average time between outages” (MTBF), or “Maintenance Free Operating Period” (MFOP) as suggested by @JDRoberts.

6 Likes

MFOP can also be an engineering target.

1 Like

I think a better question would be: daddy, are we there yet? :wink:

7 Likes

Yup…

But the severity of the outages / maintenance is also important to track for trends.

The duration, number of users, and amount of functionality affected, etc., all play a part in measuring severity.

1 Like

Lol why are you asking new questions when they can’t even answer the promised one:

what happened?

6 Likes

Or: Why do I subject myself to this?

2 Likes

I’m come from a enterprise background and we absolutely have specific target level of reliability set to specific dates. This is especially true for any revenue impacting services where five nines are a minimum.

Your point is taken and I agree with you. It’s my fault for trying to apply enterprise level expectation to a consumer level service like SmartThing but I was shocked to see that there’s NEVER been a single trouble free month in recent history.

1 Like

So… today is over… yesterday is gone… pretty sure early tomorrow happened already…

still waiting for that “deeper” explanation.

7 Likes

You may be waiting on it, but I know you didn’t expect it.

Honestly, I am hoping we stop reminding ST about all the missed commitments on the weeklys and information commits like this… just let them slide by for weeks/months, etc and then compile the list of failures.

1 Like