Closed armin-malek closed 6 months ago
Hello @armin-malek ,
Thanks for your issue.
I never used this provider so I can't say if it is coming from this or not. But I have other ideas in mind. Can you answer those questions please ?:
Regards,
we have close to 8 million files in the bucket and we serve around 7 objects per second (in reality it's a burst of 30 to 40 objects every few seconds), the average object size is around 70KB
I think I have an idea of why.
People requesting different files all the time ? Or are they the same for multiple minutes ?
it's like an image hosting platform so yeah it should be mostly random objects
Ok so I think I know why. This must come from prometheus metrics which are saving metrics per "request path". This can be patched for sure.
Ok so I think I know why. This must come from prometheus metrics which are saving metrics per "request path". This can be patched for sure.
I'm not familiar with the code base but maybe an option to disable the metrics in the config file should do the trick for now, i'll keep restarting the container every once in a while :)
I've just released a new version. The 4.13.0 release have a configuration dedicated for this.
Here is an example:
# Metrics configuration
metrics:
# Disable router path save in HTTP metrics
disableRouterPath: true
That should fix your problem.
PS: Out of this, do you use s3-proxy to make a tile map server ?
thanks, it has solved the problem!
PS: Out of this, do you use s3-proxy to make a tile map server ? no it is just serving images from a private bucket for a machine learning project
out of this too currently it's logging every request to stdout, is there way to disable that too? i couldn't find it in the docs if i'm not mistaken writing to stdout is not exactly cheap what other log levels we have in the config file?
Cool to hear this ! That's great!
Yes that's a feature to log everything. You can change the log level to error if you want only errors. All log levels from uber/zap logger are supported.
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 10 days
after startup, it slowly starts using more memory, in a few hours it reaches over 1GB because it works just fine with much less memory usage at startup, I think it's caused by a memory leak
To Reproduce here is the config file I'm using my provider is Idrive e2
Expected behavior less and stable ram usage
Version and platform (please complete the following information):