See tritonmedia. It was moved to get out of a mono-repo and be better organized.
A pipeline to allow media processing.
- API - Allows you to view statistics about the media, i.e processing status
- Deploy - Handles orchestration of Digital Ocean instances
- Events - Queues media and orchestrates new media into the pipeline
- Converter - Converts media (may change w/ deploy)
- Twilight - Organizes and deploys new media. Runs on storage host.
- Beholder - Watches over the platform (metrics and etc)
THIS IS LOSELY WRITTEN AND MAY NOT REFLECT FINAL PRODUCT
- Pull the repo down on a server
git clone <url> --recursive
- Edit
config/config.yaml
to reflect various keys / server locations.
docker-compose up
๐
If you want to use this in production I highly recommend you setup a GCP instance or a rancher instance in the cloud with Kubernetes.
Please configure the following labels on your nodes / rancher hosts:
triton.role: bigdata (for converter)
triton.role: storage (for whatever machine has the emby volume on it)
triton.state: stable (for whatever machine has internet)
Modify .ops/values.yaml
to reflect the external IP you have.
Then install our helm chart: cd .ops; helm install --name media media-stack
BSD-3-Clause