[RESOLVED] Sometimes slow on Gladys Plus dashboard

Configuration
41

For now, no, but error messages did not necessarily arrive immediately, so I will wait a few days.

42

Currently, I have about 120 ping at the moment (on the gateway), but no Gladys updates yet.

Edit: 20 ping locally and between 100 and 140 on the gateway (version 4.0.7 :slight_smile: )

43

Hi @pierre-gilles,

After a few days of trouble-free operation, I unfortunately have the error messages back on the 4 boxes of my dashboard as in my message from 01/28. The ping has not necessarily increased, but the loading of the boxes is slow again and ends with an error.

I remain available if you want to take a look at the logs.

44

Hi @Mastho!

Bummer. In my opinion, there must be an integration on your end that is causing issues. From what we’ve seen, it’s clearly a « local Â» problem with the instance.

What integrations do you already have configured? And with what settings?

Could you possibly send me your local database in private so I can try running your config locally on my end? (If you don’t mind and there’s nothing sensitive in it)

If you don’t mind, the database is located on the Pi at:

/var/lib/gladysassistant/

There are 3 files I need (gladys-production.db, gladys-production.db-shm, gladys-development.db-wal). If you can stop Gladys before copying the files, that’s better :slight_smile:

Otherwise, we can call tomorrow and investigate further what’s not working on your instance.

45

I have OpenWeather, Telegram, two cameras, and Owntracks that we had deactivated together. I still haven’t reactivated it since.

Yes, I can send you all that without any issues. I’ll retrieve it and transfer it to you as soon as possible.

Today I’m not available, but we can schedule a session in DM whenever.

46

Top :slight_smile: Send me the DB first and I’ll see what I can find with it!

47

@Mastho Well, on your DB side I didn’t find anything special.

The fact that it takes time to appear as a bug makes me think of an option: a memory leak. Maybe there is a memory leak (I’m especially thinking about the cameras), which would cause the RAM usage to increase, increase, increase until your Pi’s RAM is full..

The problem with this theory is that you should also see problems locally, but here it’s only remotely


Have you already checked the RAM usage on your Pi when you had the bugs (Type htop in the command line)?

Otherwise, I would need all the logs from your containers to see if something special is happening on your side.

For this, you can do on your Pi:

docker logs gladys > logfile.log 2>&1

And send me the log file :slight_smile: (by email if you want)

We will find it!! :smiley:

48

I monitored the RAM and it remains at 25% usage, whether Gladys processes a request or not.
I will check again when the dashboard error messages reappear. They disappeared after the last restart.

I just sent you the log. Happy searching :slight_smile:

49

So the log you sent me is empty, since Gladys updated in between (to v4.1.0), which renewed the container and cleared the logs ^^

The best thing is that as soon as you have the problem again:

  • You directly export the logs with the command I gave you
  • You run htop and take a screenshot of your console

You send me both, and I’ll take a look :slight_smile: