#Extremely slow

26 messages · Page 1 of 1 (latest)

manic mesa
#

My seer is often extremely slow.

Symptoms:

  • Discover/Movies/Series pages are very slow loading
  • Images of items are very slow to load
  • When you click on an item (a movie/serie), loading is even slower (60+ seconds)
  • Settings page is extremely fast
  • Other apps like Sonarr and Radarr are extremely fast

Setup:

  • Docker compose image
  • LXC running on a m2 HDD
  • LXC is on a 192.x internal IP address
  • My 10.x LAN is proxied via a Caddy to jellyseer
x-common-env: &common-env
  PUID: "1000"
  PGID: "10000"
  TZ: "Europe/Amsterdam"
  LOG_LEVEL: debug
  UMASK: "002"

services:
  seerr:
    image: ghcr.io/fallenbagel/jellyseerr:latest
    container_name: seerr
    environment:
      <<: *common-env
      LOG_LEVEL: info
    volumes:
      - /root/docker/seerr:/app/config
    ports:
      - "5055:5055"
    restart: unless-stopped
    networks:
      - arr_net
    healthcheck:
      test: wget --no-verbose --tries=1 --spider http://localhost:5055/api/v1/status || exit 1
      start_period: 20s
      timeout: 3s
      interval: 15s
      retries: 3

Already tried:

  • Changed the LXC to not use SWAP memory
  • Enabled Image Caching (since then images loaded are a bit faster, but uncached images still 30+ seconds)
#

What helps to debug this? Logs, if so which? Browser logs?

manic mesa
#

docker logs -f seerr

2026-02-12T17:49:23.138Z [info][API]: Found matching Jellyfin user; updating user with Jellyfin {"ip":"10.0.0.10","jellyfinUsername":"Jeroen"}
2026-02-12T17:50:00.002Z [info][Jobs]: Starting scheduled job: Jellyfin Recently Added Scan
2026-02-12T17:50:00.003Z [info][Jellyfin Sync]: Jellyfin Sync Starting {"sessionId":"ea4527d9-dd64-4185-b56c-99b0940a04c1"}
2026-02-12T17:50:00.005Z [info][Jellyfin Sync]: Beginning to process recently added for library: Movies

First line appears when I open the app in the browser. But even with many items having "blank images", no logs appear.

After a while, these Jobs/Jellyfin items appear. (images were still loading)

77250045-DC51-4B05-AC44-144A931B6F33.png
#

After around a minute ( I can time it exactly if you need it), the images appeared in the browser.

next steppe
#

Yeah thats networking/dns related

#

Evidenced by "settings page loads extremely fast"

#

The cards you see on those pages rely on dns/networking

#

So I would recommend adding these to your environment

# 
dns:
  - 9.9.9.9
  - 8.8.8.8
  - 1.1.1.1
#

Then on settings > network > enable forceIpv4First option

#

And restart

#

And test now

#

If its still slow there are some other few things that can be done

manic mesa
#

oooh, it IS dns

# 
root@arr:~# docker exec -it seerr sh -lc '
echo "START $(date -Iseconds) hostname"
time wget -S -T 20 -O /dev/null "http://jellyfin-server.internal.blahbla.com:8096/System/Info/Public" >/dev/null 2>&1
echo "END   $(date -Iseconds)"

echo
echo "START $(date -Iseconds) IP"
time wget -S -T 20 -O /dev/null "http://192.168.0.27:8096/System/Info/Public" >/dev/null 2>&1
echo "END   $(date -Iseconds)"
'
START 2026-02-12T20:46:47+01:00 hostname
END   2026-02-12T20:46:51+01:00

START 2026-02-12T20:46:51+01:00 IP
END   2026-02-12T20:46:51+01:00
root@arr:~#
next steppe
next steppe
#

If still slow add these to compose

# 
    sysctls:
      net.ipv6.conf.all.disable_ipv6: "1"
      net.ipv6.conf.default.disable_ipv6: "1"
      net.ipv6.conf.lo.disable_ipv6: "1"
#

So the docker container doesnt use ipv6

#

Sometimes on some rare networks, ipv6 messes up nodejs's dns system

manic mesa
#

It's always DNS

  seerr:
    image: ghcr.io/fallenbagel/jellyseerr:latest
    container_name: seerr
    dns:
      - 192.168.0.2

Workaround works, everything is very fast now.

My technitium LXC has two interfaces (10.0.1.2 / 192.168.0.2). I need to check why its telling my docker LXC (who lives on 192.x) that there's a nameserver on 10.0.1.2 as well. 😄

next steppe
manic mesa
#

So not sure why jellyseerr tried to use the unreachable dns server (10.0.1.2)

manic mesa
#

Ah looked that up; I didn't know the fallback to DNS server 2 wasnt immediately, but dependent on timeouts & retries