Bonus - accidental logging of crashes!

My crashing disease got added interest today - I noted that Grafana graph of my temperatures also showed me flat spots whenever the hub had gone dead. The graph is set 'connect points when no data' which gave me flat lines whenever my C7 went bad! Cool use of technology!

Here is just the last week -

1 Like

You sure those are actual dead spots and not just spans of no change?

4 Likes

nope.
I know for a fact at least 60% of them are real - they occur while I'm around. I've had to reset since lights stop going on off and alexa says 'device not responding'.

Since I took this graph I've enjoyed 2 days of no crashes. I have been disabling apps a few at a time to see if any impact. I disabled Hub Information and ECC (power monitoring) and I've been stable for 2 days - so. for now I'm leaving it just exactly as it is until/if next crash. it may have been as simple as an app problem all along. It doesn't account for numerous other issues found - like the dhcp problem the dns problem etc. but those are overcome now.
So. we'll see. my point of the post was not really to highlight my crashes, it was to point out sometimes there are other methods to capture data.
Here's todays' 7 day graph :slight_smile:

2 Likes

Happy to report - been stable now for 10 days. I reenabled Hub Info 5 days ago. No problems.
For a test I reenabled ECC (its listed as Beta) and within 10 hours, crashed. So. found the problem I guess. I additionally replaced the 3rd reality units as well and replaced with basic outlets since they became unnecessary due to not gathering the power info any longer.
I also can explain 2 'flat spots' that happen every night - 12am and 2am. Both are OpenWeatherMap not providing data for short durations - probably website updates/backups or something.

I highly doubt my app is the root of your problems, but the extra little bit of hub load might be.

Beta just means it's not what I would consider 100% functional since there's a bit more I'd planned to do with it.

1 Like

I didn't mean to imply the app was faulty, just that it was connected to my problem. Since it relied on some shady devices, my guess is it was either the data I tableized to setup the feedback or the devices themselves.
image