Skip to content

Service Catalog: Media & Automation

This section documents the automated media indexing, hardware-accelerated transcoding pipelines, streaming platforms, and dedicated game servers operating within the LambertLab ecosystem.

🎬 Workload Directory

1. Jellyfin Media Server

  • Namespace: media
  • Sync Wave: Wave 3
  • Ingress Endpoint: https://jellyfin.lambertlab.us
  • Architectural Role: Privacy-focused, high-performance open-source media streaming server.
  • Storage Backend: Persistent application configuration and transcode cache on host NVMe (/home/chris/services/k3s-jellyfin/config on workstation, 15Gi hostPath); media libraries mounted via ReadWriteMany (RWX) NFS shares directly from the TerraMaster NAS 14TB pool (/Volume1/data).
  • Hardware Acceleration:
  • Scheduled with node affinity targeting workstation.
  • Passes through the NVIDIA RTX A4500 (20 GB VRAM) via nvidia.com/gpu: 1 resource requests.
  • Enables zero-overhead NVENC (hardware encoding) and NVDEC (hardware decoding) across 4K HDR HEVC/H.265 streams with tone-mapping.

2. Sonarr

  • Namespace: media
  • Sync Wave: Wave 3
  • Ingress Endpoint: https://sonarr.lambertlab.us
  • Architectural Role: Automated TV series collection manager and download scheduler.
  • Storage Backend: Dedicated 20 GB raw iSCSI block LUN (iqn.2026-06.us.lambertlab:sonarr-config on TerraMaster SAN) for application database and settings; ReadWriteMany NFS mount (/Volume1/data) for media library organization.
  • Key Configuration:
  • Reconciles television series monitoring, season upgrades, and file renaming.
  • Integrates with Prowlarr for upstream indexer coordination.

3. Radarr

  • Namespace: media
  • Sync Wave: Wave 3
  • Ingress Endpoint: https://radarr.lambertlab.us
  • Architectural Role: Automated movie collection manager and quality profile scheduler.
  • Storage Backend: Dedicated 20 GB raw iSCSI block LUN (iqn.2026-06.us.lambertlab:radarr-config on TerraMaster SAN) for movie database and configurations; ReadWriteMany NFS mount (/Volume1/data) for media library management.
  • Key Configuration:
  • Tracks release dates, monitors wanted films, and orchestrates automatic quality upgrades.
  • Configured with custom quality definitions and format scoring for optimal video/audio codec matching.
  • Shared NFS volume claims with Sonarr and Jellyfin for unified library management.

4. Prowlarr

  • Namespace: media
  • Sync Wave: Wave 3
  • Ingress Endpoint: https://prowlarr.lambertlab.us
  • Architectural Role: Centralized indexer proxy integrating Usenet and BitTorrent feeds.
  • Storage Backend: Dedicated 2 GB raw iSCSI block LUN (iqn.2026-06.us.lambertlab:prowlarr-config on TerraMaster SAN) for indexer SQLite database and configurations.
  • Key Configuration:
  • Manages and tests upstream tracker connections and rate limits.
  • Synchronizes indexers seamlessly across both Sonarr and Radarr via in-cluster API endpoints, eliminating redundant manual configuration.

5. Tdarr Transcoding Node

  • Host Environment: workstation (Docker Engine managed via Terraform docker/workstation/app-tdarr.tf)
  • Sync Wave: Standalone Host Engine
  • Ingress Endpoint: https://tdarr.lambertlab.us
  • Architectural Role: Distributed, automated media library transcoder and audio/video standardization pipeline.
  • Key Configuration:
  • Mounts the host NVIDIA GPU device (/dev/dri and NVIDIA runtime) for batch transcoding video libraries to space-efficient AV1 or HEVC formats.
  • Directly accesses the high-throughput 14TB TerraMaster NFS pool for line-rate read/write processing.
  • Runs as a containerized worker communicating with the central Tdarr server coordinator.

6. Palworld Dedicated Game Server

  • Namespace: gaming
  • AppProject: gaming
  • Sync Wave: Wave 3
  • Network Port: Port 8211 UDP
  • Architectural Role: Dedicated, high-performance multi-player game server.
  • Key Configuration:
  • Pod pinned deterministically to opti74 via nodeSelector: kubernetes.io/hostname: opti74 to isolate gaming memory consumption from core manager workloads.
  • Backed by a retained Longhorn persistent volume claim (palworld-data) to guarantee player progression and world state survive cluster sync waves.
  • Configured with memory limits (limits.memory: 12Gi) and auto-restart policies to manage game server memory leaks.