Aletheia
Log in
Operations

Deployment Guide

Production deployment should preserve durability first, then optimize for latency and throughput.

Deployment topology

A common topology runs Aletheia as a dedicated memory service behind an internal API gateway. Keep data directories on persistent volumes with regular backups.

Avoid ephemeral disks for primary data unless you have robust replication and recovery strategy.

Environment checklist

  • Persistent volume for redb and index data
  • Model cache directory with predictable permissions
  • Resource limits sized for reranking workloads
  • Readiness and liveness probes
  • Structured log export and metrics scraping

Container starter

Example run
docker run --rm -p 3000:3000 \
  -v /srv/aletheia-data:/data \
  -e ALETHEIA_DATA_DIR=/data \
  ghcr.io/aletheia/aletheia:latest