compose.yml now holds each default as ${VAR:-default}, and the code keeps
no fallback values. The bot stops at startup on a missing or invalid
PORT or LOG_LEVEL, /addsticker refuses without STICKER_PACK_NAME, and the
renderer refuses to start until every RENDERER_* setting is a valid
value, listing each problem. Settings whose empty value means none or
all (MODULES, OWNER_ID, ADMIN_IDS, the API tokens) use ${VAR:-}.
The renderer's npm start and dev load .env when present, so a local run
works from a copy of .env.example.
BREAKING CHANGE: running outside compose now requires LOG_LEVEL and PORT
for the bot, STICKER_PACK_NAME for /addsticker, and every RENDERER_*
variable for the renderer.
2.1 KiB
Deployment
miti99bot compose
The root compose.yml builds this folder as the renderer service and points
the bot at it over the compose network. The API has no authentication: it is
reachable only from inside that network, so never publish its port or attach a
domain to it. The rest of this page covers running the service on its own.
Recommendation
Use a self-hosted/container runtime for v1. Static-only hosts are not enough
because /api/gif must render GIF bytes on the server with Remotion.
Good first targets:
- Coolify Docker app
- Fly.io
- Railway
- Render
- VPS with Docker
- Google Cloud Run generic container
Avoid for v1:
- Cloudflare Workers/Pages
- pure static Vercel/Netlify deploys
Runtime
Env vars, all required (standard values shown):
RENDERER_PORT=3000
RENDERER_HOST=0.0.0.0
RENDERER_MAX_CONCURRENT_RENDERS=1
RENDERER_RENDER_TIMEOUT_MS=15000
RENDERER_MAX_OPTIONS=32
RENDERER_MAX_OPTION_CHARS=40
The renderer has no fallback values: it refuses to start and lists every
missing or invalid variable. The root compose.yml owns the defaults — it
fixes RENDERER_HOST and RENDERER_PORT and gives each tuning value a
${VAR:-default}, so leaving them empty in Coolify uses these values. For a
local run, copy .env.example to .env; npm run dev and npm start load it.
Start with 1-2 vCPU and 1-2 GB RAM. Increase only after render benchmarks show the service is CPU-bound or concurrency-limited.
Local non-Docker render smoke requires Chrome Headless Shell shared libraries,
including libnspr4 and libnss3. Prefer Docker for consistent verification.
The Docker image runs npm run browser:ensure during build so production requests
do not need to download Chrome Headless Shell on first render.
RENDERER_RENDER_TIMEOUT_MS is a total render timeout. Values below 7000 are raised to
7000 because Remotion's browser timeout has that minimum.
Health
curl http://localhost:3000/api/healthz
Render Test
curl -X POST http://localhost:3000/api/gif \
-H 'content-type: application/json' \
--output wheel.gif \
--data '{"options":["alice","bob","carol"],"winnerIndex":1}'