[Bug]: Memory Leak on Docker Deployment

[nerdc0re-net]

Description

Since i have Bulwark deployed via docker-compose.yaml on my Dockhand infrastrucutre I am able to observe that Bulwark has some kind of memory leak. I would not expect a webmail client to eat up endless memory up to 2 GB just for the container (or even more depending on the host machines max memory).

I have to mention that I do not limit resources on my containers if not necessary, so there is probably some way to circumvent this issue via that route, but I think this shouldn't be the expected behaviour nonetheless.

Steps to Reproduce

1. start Bulwark container
2. watch memory usage constantly rise on the host machine

Expected Behavior

no endless memory consumption for a simple webmail client.

Actual Behavior

memory leak.

Bulwark Version

1.4.11

Stalwart Mail Server Version

0.15

Browser

Chrome / Chromium

Operating System

Linux

Screenshots / Screen Recording

Relevant Logs or Error Output

Additional Context

I added the memory usage of my host machine, every time the ramp is rising again it's because I restarted the container, last flat-line was a longer duration shutdown before starting Bulwark again.

[chrilep]

Sorry to see that. I never experienced this.

Could you please share redacted info about your container's settings?

[rathlinus]

Yes please share more info, my container uses about 50-70mb

[rathlinus]

Please also paste this: https://webmail.example.com/api/health?detailed=true

[nerdc0re-net]

So my container currently has memory-usage around 919MB (according to Dockhand), according to docker via CLI:

CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
************ bulwark 0.00% 929.7MiB / 4GiB 22.70% 1.26MB / 3.9MB 23.7MB / 0B 11

and my docker-compose looks like this:

bulwark:
image: ghcr.io/bulwarkmail/webmail:latest
container_name: bulwark
ports:
- "3003:3000"
environment:
- TZ=Europe/Vienna
- JMAP_SERVER_URL=****
- OAUTH_ENABLED=true
- OAUTH_CLIENT_ID=${OAUTH_CLIENT_ID}
- OAUTH_CLIENT_SECRET=${OAUTH_CLIENT_SECRET}
- OAUTH_ISSUER_URL=****
depends_on:
- stalwart
healthcheck:
test:
[
"CMD",
"wget",
"--no-verbose",
"--tries=1",
"--spider",
"http://127.0.0.1:3000/api/health",
]
interval: 30s
timeout: 5s
retries: 3
start_period: 10s
restart: unless-stopped

And the output you were asking for is here:

{"status":"healthy","timestamp":"2026-04-07T13:58:36.503Z","uptime":11229.746638854,"version":"0.1.0","memory":{"heapUsed":114127296,"heapTotal":123158528,"heapSizeLimit":4496293888,"rss":953831424,"external":82108877,"heapUsagePercent":2.54},"environment":"production","nodeVersion":"v24.14.1"}

[chrilep]

You are on latest. Maybe give latest main a try. Bulwark gets tons of fixes and features every day.

Currently "my" Bulwark over here has these details:

{
  "status": "healthy",
  "timestamp": "2026-04-07T14:41:53.858Z",
  "uptime": 18095.34836134,
  "version": "0.1.0",
  "memory": {
    "heapUsed": 60711808,
    "heapTotal": 63741952,
    "heapSizeLimit": 2348810240,
    "rss": 139395072,
    "external": 9074585,
    "heapUsagePercent": 2.58
  },
  "environment": "production",
  "nodeVersion": "v24.14.1"
}

and docker...

> docker container stats bulwark

CONTAINER ID   NAME      CPU %     MEM USAGE / LIMIT     MEM %     NET I/O           BLOCK I/O    PIDS
e5ea51454703   bulwark   0.00%     97.98MiB / 7.607GiB   1.26%     2.01MB / 2.78MB   320MB / 0B   11

The WebUI shows no RAM changes at all:

Additional to using latest main image you could set more ENV vars as given in the example env file and the docs.

• https://github.com/bulwarkmail/webmail/blob/main/.env.example
• https://bulwarkmail.org/docs/getting-started/configuration/environment-reference

@rathlinus may have additional ideas...

[rathlinus]

Have you been uploading lots of files?

[nerdc0re-net]

Have you been uploading lots of files?

Nope, I already saw the open issue concerning the file uploads, but thats not the case. This was simply a deployment for testing purposes. No uploads happened at all here. Once I'm back home tomorrow I gonna check out the environment files and give latest main a try to see how that turns out.

[rathlinus]

Sounds good

[nerdc0re-net]

Hey again! So I switched to main, added some of the env variables, even stripped most of them like OIDC-related ones to check if this could be related to the issue, but I got this memory leak on all of these different setups.

It takes around 2 hours until the container reaches 512M memory usage... at this point I simply limit the mem and swap usage.

[rathlinus]

I will still try to replicate it, but i was not able to yet