Penpot crashes when saving files

Self-host Installation
,
1

Hi!

I installed a Penpot instance via Docker on a Debian VM last November and everything was working perfectly.

Not knowing anything about Docker, I couldn’t update it correctly and had to reinstall everything in February. I used the official version of Docker and installed it on a specific /data partition to avoid saturating the system disk. Since then, I’ve had constant problems and Penpot has become very unstable:

Client side: (Firefox or Chrome)

Even small projects seem to take a long time to load completely, and update files take longer and longer (blue “saving” dot remains) and eventually crash (page “internal error,” status 504, code: repository-access-error).

At the same time, wss requests accumulate in the browser and increasingly return timeout errors.

Server side:

CPU consumption seems enormous (120% on average for 2 CPUs, while RAM consumption is less than 20%), even when Penpot is not open.

Penpot’s logging system seems to be the source of this overload.

The backend logs are saturated with continuous BLPOP redis timeouts.

I’ve tried a lot of things: increasing the nginx and valkey timeouts, manually activating the workers via docker-compose (which generated a lot of zombie processes), deactivating them (they still appear to be activated, but I only have one zombie process left).

I followed the installation instructions carefully, checked my URI, and configured my nginx WAF (front-end on another wm and shared for other services) as recommended.

If I restart everything, it works again, but it degrades within a few hours even if I don’t connect to Penpot.

I upgraded Penpot to the latest version available, and nothing changes no matter what I do.

I really want to use Penpot for my projects, but for now it’s unmanageable.

Does anyone have any ideas?

Thanks in advance to anyone who can help me out.

2

Have you looked at the Redis logs to see if they explain the timeout errors?

3

This all I have on valkey logs since my last restart

1:M 18 Aug 2025 14:06:58.361 * oO0OoO0OoO0Oo Valkey is starting oO0OoO0OoO0Oo
1:M 18 Aug 2025 14:06:58.361 * Valkey version=8.1.3, bits=64, commit=00000000, m odified=0, pid=1, just started
1:M 18 Aug 2025 14:06:58.361 # Warning: no config file specified, using the defa ult config. In order to specify a config file use valkey-server /path/to/valkey. conf
1:M 18 Aug 2025 14:06:58.361 * monotonic clock: POSIX clock_gettime
1:M 18 Aug 2025 14:06:58.361 * Running mode=standalone, port=6379.
1:M 18 Aug 2025 14:06:58.361 * Server initialized
1:M 18 Aug 2025 14:06:58.361 * Ready to accept connections tcp
1:M 19 Aug 2025 02:01:36.935 * 1 changes in 3600 seconds. Saving…
1:M 19 Aug 2025 02:01:36.935 * Background saving started by pid 326764
326764:C 19 Aug 2025 02:01:37.045 * DB saved on disk
326764:C 19 Aug 2025 02:01:37.045 * Fork CoW for RDB: current 0 MB, peak 0 MB, a verage 0 MB
1:M 19 Aug 2025 02:01:37.136 * Background saving terminated with success
1:M 19 Aug 2025 07:16:40.286 * 1 changes in 3600 seconds. Saving…
1:M 19 Aug 2025 07:16:40.286 * Background saving started by pid 470088
470088:C 19 Aug 2025 07:16:40.347 * DB saved on disk
470088:C 19 Aug 2025 07:16:40.348 * Fork CoW for RDB: current 0 MB, peak 0 MB, a verage 0 MB
1:M 19 Aug 2025 07:16:40.386 * Background saving terminated with success
1:M 19 Aug 2025 08:53:27.899 * 1 changes in 3600 seconds. Saving…
1:M 19 Aug 2025 08:53:27.899 * Background saving started by pid 472327
472327:C 19 Aug 2025 07:21:35.869 * DB saved on disk
472327:C 19 Aug 2025 07:21:35.869 * Fork CoW for RDB: current 0 MB, peak 0 MB, a verage 0 MB
1:M 19 Aug 2025 07:21:35.933 * Background saving terminated with success

4

But i have plenty of that in backend logs

[2025-08-19 11:32:27.912] E app.worker.runner - hint="redis pop operation timeout, consider increasing redis timeout (will retry in some instants)", timeout=#app/duration "5s"
SUMMARY:
 →  io.lettuce.core.RedisCommandTimeoutException: Command timed out after 10 second(s) (ExceptionFactory.java:63)
DETAIL:
 →  io.lettuce.core.RedisCommandTimeoutException: Command timed out after 10 second(s) (ExceptionFactory.java:63)
    at: io.lettuce.core.internal.ExceptionFactory.createTimeoutException(ExceptionFactory.java:63)
        io.lettuce.core.internal.Futures.awaitOrCancel(Futures.java:233)
        io.lettuce.core.FutureSyncInvocationHandler.handleInvocation(FutureSyncInvocationHandler.java:79)
        io.lettuce.core.internal.AbstractInvocationHandler.invoke(AbstractInvocationHandler.java:86)
        jdk.proxy2.$Proxy32.blpop(:-1)
        app.redis$initialize_resources$reify$reify__24059.blpop(redis.clj:261)
        app.worker.runner$run_worker_loop_BANG_.invokeStatic(runner.clj:230)
        app.worker.runner$run_worker_loop_BANG_.invoke(runner.clj:151)
        app.worker.runner$start_thread_BANG_$fn__58548.invoke(runner.clj:263)
        promesa.exec$binding_conveyor_inner$fn__13480.invoke(exec.cljc:193)
        clojure.lang.AFn.run(AFn.java:22)
        java.lang.Thread.run(:-1)
5

Have you tried connecting to Penpot directly on the VM port without going through your nginx proxy to see if the 504 errors still happen?

6

Yes and that’s the same.

7

Hmmm… could it be some time mismatch between the host and the containers?

8

Maybe, I can see the container are using UTC and my host is using CEST.
Can that be the problem ?