s-schwarzmann/docker-video-encoding

docker-video-encoding's Issues

Collect information about RAM usage

Stackoverflow

Change to docker compose

Add reference video to calculate ssim, psnr and vmaf

Predict/Monitor needed ressources

Running encoding and analyzing videos need a lot of resources:

• RAM
• Disk-Storage

ffmpeg has different failure behaviors.
If a machine runs out of ram and is switching to swap, weird encoding will happen. ffmpeg won't report an error. That is why some ram monitoring would be great and to have the ability to freeze some jobs.

Videos need a lot of disk-storage. It would be great if a job is only started if enough disk space is available. Running out of disk storage will result in a ffmpeg error. So bad encoding can be detected.

sftp library fails

Uploading huge files (4k Videos) can result in session renegotiation failures and starvation of the process

paramiko.ssh_exception.SSHException: Key-exchange timed out waiting for key negotiation

include name of machine if job fails

Currently, if a job fails u have no idea where the job was running.
Jobs that are running have the name of the vm in the machine.

Compare:
https://github.com/s-schwarzmann/docker-video-encoding/blob/master/dirjobs.py#L92
to:
https://github.com/s-schwarzmann/docker-video-encoding/blob/master/dirjobs.py#L134

I think it is a nice idea to include such information.

Rename .txt to .json

Why are the job files .txt and not .json?
IDEs would directly recognize that the job files are json-formatted if you would use .json. :)