Docker / DB: FATAL: sorry, too many clients already

[elegos]

Describe the issue
I installed MediaCMS yesterday evening and uploaded a couple of videos. This morning the web, celery_worker and celery_beat services were already offline. The logs showed nothing fancy (also because docker compose cycled the logs after the restart), but I noticed the "FATAL: sorry, too many clients already" error in the db log spamming for roughly 5:20 hours before shutting down (and I think letting the above services crash).

To Reproduce
Steps to reproduce the issue:

1. Run via docker compose
2. Upload a couple of videos
3. Let them transcribe
4. Watch the videos a couple of times

Expected behavior
No crash.

Screenshots
N-A

Environment (please complete the following information):

• OS: Docker 28.5.0 // Compose 2.39.4 @ Fedora Linux 42
• Installation method: Docker via docker compose
• Browser, if applicable N/A

Additional context
Last lines of spam where:

db-1  | 2025-10-07 05:22:17.660 BST [15434] FATAL:  sorry, too many clients already
db-1  | 2025-10-07 05:22:27.715 BST [15441] FATAL:  sorry, too many clients already
db-1  | 2025-10-07 05:22:37.764 BST [15449] FATAL:  sorry, too many clients already
db-1  | 2025-10-07 05:22:47.811 BST [15456] FATAL:  sorry, too many clients already
db-1  | 2025-10-07 05:22:57.867 BST [15464] FATAL:  sorry, too many clients already
db-1  | 2025-10-07 05:22:59.894 BST [1] LOG:  received fast shutdown request
db-1  | 2025-10-07 05:22:59.903 BST [1] LOG:  aborting any active transactions
db-1  | 2025-10-07 05:22:59.903 BST [14893] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [14696] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13927] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [14720] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13703] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13793] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13794] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [11996] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13688] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.903 BST [13685] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.904 BST [14695] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.904 BST [13795] FATAL:  terminating connection due to administrator command

...

db-1  | 2025-10-07 05:22:59.922 BST [2217] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.922 BST [2252] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.925 BST [2193] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.930 BST [1] LOG:  background worker "logical replication launcher" (PID 32) exited with exit code 1
db-1  | 2025-10-07 05:22:59.932 BST [2177] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.934 BST [2173] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.933 BST [2196] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.917 BST [2174] FATAL:  terminating connection due to administrator command
db-1  | 2025-10-07 05:22:59.979 BST [27] LOG:  shutting down
db-1  | 2025-10-07 05:22:59.986 BST [27] LOG:  checkpoint starting: shutdown immediate
db-1  | 2025-10-07 05:23:00.034 BST [27] LOG:  checkpoint complete: wrote 1 buffers (0.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.003 s, sync=0.010 s, total=0.056 s; sync files=1, longest=0.010 s, average=0.010 s; distance=0 kB, estimate=171 kB; lsn=0/1CD9B40, redo lsn=0/1CD9B40
db-1  | 2025-10-07 05:23:00.042 BST [1] LOG:  database system is shut down

[mgogoulos]

it seems that the app is opening too many connections or not releasing them properly / haven't seen anything similar before. What are the server specs, is there any constraint there? Also was there any traffic on the system while this happened?

[elegos]

Hello @mgogoulos,

Specs are (from fastfetch):

         .';:cccccccccccc:;,.             --------------
      .;cccccccccccccccccccccc;.          OS: Fedora Linux 42 (Server Edition) x86_64
    .:cccccccccccccccccccccccccc:.        Host: EffiZen Series (Version 1.0)
  .;ccccccccccccc;.:dddl:.;ccccccc;.      Kernel: Linux 6.16.10-200.fc42.x86_64
 .:ccccccccccccc;OWMKOOXMWd;ccccccc:.     Uptime: 8 hours, 47 mins
.:ccccccccccccc;KMMc;cc;xMMc;ccccccc:.    Packages: 1105 (rpm)
,cccccccccccccc;MMM.;cc;;WW:;cccccccc,    Shell: zsh 5.9
:cccccccccccccc;MMM.;cccccccccccccccc:    Terminal: /dev/pts/0
:ccccccc;oxOOOo;MMM000k.;cccccccccccc:    CPU: AMD Ryzen 7 5800U (16) @ 4.51 GHz
cccccc;0MMKxdd:;MMMkddc.;cccccccccccc;    GPU: AMD Radeon Vega Series / Radeon Vega Mobile Series [Integrated]
ccccc;XMO';cccc;MMM.;cccccccccccccccc'    Memory: 17.22 GiB / 28.21 GiB (61%)
ccccc;MMo;ccccc;MMW.;ccccccccccccccc;     Swap: 12.00 KiB / 23.96 GiB (0%)
ccccc;0MNc.ccc.xMMd;ccccccccccccccc;      Disk (/): 244.48 GiB / 935.53 GiB (26%) - xfs
cccccc;dNMWXXXWM0:;cccccccccccccc:,       Local IP (enp2s0): 192.168.1.10/24
cccccccc;.:odl:.;cccccccccccccc:,.        Locale: it_IT.UTF-8
ccccccccccccccccccccccccccccc:'.
:ccccccccccccccccccccccc:;,..                                     
 ':cccccccccccccccc::;,.                                          

So I think there are no particular constraints. There should have been 0 or almost 0 traffic at that time: the dns entry, reverse proxy and the whole mediacms service where just have been setup, and it was late in the evening.

Since that event, the system is working "just fine" apparently:

gfurlan@helium ~/docker/mediacms main* ❯ docker compose ps                                                                                                                                                                          15:28:07
WARN[0000] /home/gfurlan/docker/mediacms/docker-compose.yaml: the attribute `version` is obsolete, it will be ignored, please remove it to avoid potential confusion 
NAME                       IMAGE                      COMMAND                  SERVICE         CREATED        STATUS                 PORTS
mediacms-celery_beat-1     mediacms/mediacms:latest   "./deploy/docker/ent…"   celery_beat     30 hours ago   Up 9 hours             80/tcp, 9000/tcp
mediacms-celery_worker-1   mediacms/mediacms:latest   "./deploy/docker/ent…"   celery_worker   30 hours ago   Up 9 hours             80/tcp, 9000/tcp
mediacms-db-1              postgres:17.2-alpine       "docker-entrypoint.s…"   db              39 hours ago   Up 9 hours (healthy)   5432/tcp
mediacms-redis-1           redis:alpine               "docker-entrypoint.s…"   redis           39 hours ago   Up 9 hours (healthy)   6379/tcp
mediacms-web-1             mediacms/mediacms:latest   "./deploy/docker/ent…"   web             30 hours ago   Up 9 hours             9000/tcp, 0.0.0.0:12380->80/tcp, [::]:12380->80/tcp

[elegos]

I've eventually managed to reproduce the problem.
As previously written, I uploaded a couple of videos (nothing fancy, roughly 15 minutes videos). One got encoded, the other one kept being in "Pending" state.

Today I tried to follow the hints in this post: #74 (comment) and once it started trying encoding the video again, it started spamming the connections issue again: FATAL: sorry, too many clients already.