Container starts ramping up CPU & mem usage

[Manuel5cc]

I had this container running for several days and it works much better than the original flaresolver, but if left running "unattended", it starts eating up resources.

These graphs show the cpu and memory evolution until I stopped it around 7 am, where you can clearly see that the resources are freed.

[ThePhaseless]

Hi, are you sure that you're using the latest version? There was a moment where a bug prevented the browser to close after failure

[Manuel5cc]

Hi, are you sure that you're using the latest version? There was a moment where a bug prevented the browser to close after failure

I pulled the :latest version so it should.

[ThePhaseless]

Hi, I've added a hotfix that ends browser on every exception, should fix your problem, please pull now

[Manuel5cc]

Hi there!
This started happening again since a few days ago. I normally pull new images on saturday so I guess if there was a new image in the last 15 days that's causing the issue.

[ThePhaseless]

That is strange... After the rework, now the implementation implies that everything closes after the request is finished (whether it was successful or not), but I've added a driver quit anyways. Please check if version v1.0.4 fixes this problem

[Manuel5cc]

That is strange... After the rework, now the implementation implies that everything closes after the request is finished (whether it was successful or not), but I've added a driver quit anyways. Please check if version v1.0.4 fixes this problem

It has started happening again after 8 hours running with the new release. The logs don't show anything out of the ordinary happening at the same time or during the high CPU periods. For now I found a workaround, I restart the container every hour.

[ThePhaseless]

Hi, I think I know what's happening. Sometimes chrome can crash at startup (I think). So there is no way to directly catch this. I've added a function that will kill all chrome processes that are longer than 5 minutes at health check endpoint (which is triggered with docker healthcheck each 5 mins). Currently this is only on :latest tag. I will create a release after testing

[ThePhaseless]

As I do not see a problem on my site anymore and no responses here, I'll close this