Possible Memory Leak?

[surjithctly]

I'm using PostgresDB and it seems that its using 100% resources all the time. I restarted it once and it went to normal, but then after few days it became full. See screenshot. Is this might have happened because of memory leaks?

Let me know what info you need to debug, like some SQL commands, I can do that.

See how memory & network is using the max available. My websites are not went viral, I have added 5 sites, and gets a maximum of 2k visits per day.

[ruibaby]

same here #1134

[dipakparmar]

@surjithctly How many connections you're having on postgres from umami?

[surjithctly]

@dipakparmar How can I check them? Only one umami site is connected.

[dipakparmar]

@dipakparmar How can I check them? Only one umami site is connected.

@surjithctly you could check by running this query or if you have pgadmin installed it will show you:

select pid as process_id, 
       usename as username, 
       datname as database_name, 
       client_addr as client_address, 
       application_name,
       backend_start,
       state,
       state_change
from pg_stat_activity;

[surjithctly]

Here are the results: @dipakparmar

[dipakparmar]

seems normal number connections! is the usage that you showed on earlier screenshot is during peak time or regular?

[surjithctly]

It's regular. Drops to normal after a restart but picks up to max in a day or two.

[surjithctly]

I'm using railway.app

[dipakparmar]

In that case, check with Railway support. You could also try running the DB on docker local and test with simulated traffic.

[surjithctly]

okay. will check with them. Not a docker expert :)

[surjithctly]

If anyone using Railway (recommended by Umami) as the database, please share your metrics screen! I wanted to see if its just me.

Coz its not solved yet.

[taylor-stein]

Unsure if it's related, but I'm seeing memory constantly rise until the server crashes, or until the deployment is automatically restarted. Followed the docs for configuring Umami with Railway. You can see the crash at around 6 AM:

[rylax]

Any update on this? Have you figured out the root cause? I am experiencing the exact same.

[surjithctly]

I had the issue while importing data from my old sql server. I created a new server without backup or import. just fresh. Then it worked as expected

[rylax]

Hm, I see. I started everything fresh and this keeps on happening unfortunately. The only logs I get are TypeError: Cannot destructure property 'session_id' of 'session' as it is null but not sure they are even related with the leak.

[dipakparmar]

@rylax try running failed migrations manually if you have any failed. I had similar problems after the upgrade and had to run migrations manually, unfortunately automatically migration did not work properly.