refactor(renderer)!: prefix every renderer env var with RENDERER_

The renderer's settings now read RENDERER_HOST, RENDERER_PORT,
RENDERER_MAX_CONCURRENT_RENDERS, RENDERER_RENDER_TIMEOUT_MS,
RENDERER_MAX_OPTIONS, and RENDERER_MAX_OPTION_CHARS, so they read as
renderer settings and cannot clash with the bot's variables, which
Coolify injects into every service. NODE_ENV keeps its standard name.

BREAKING CHANGE: the unprefixed HOST, PORT, MAX_CONCURRENT_RENDERS,
RENDER_TIMEOUT_MS, MAX_OPTIONS, and MAX_OPTION_CHARS are no longer read
by the renderer.

compose.yml

@@ -57,12 +57,14 @@ services:
     restart: unless-stopped
     environment:
       NODE_ENV: production
-      HOST: 0.0.0.0
+      # Every renderer setting carries the RENDERER_ prefix so it cannot clash
-      PORT: "3000"
+      # with the bot's variables, which Coolify also injects into this service.
-      MAX_CONCURRENT_RENDERS: ${MAX_CONCURRENT_RENDERS:-1}
+      RENDERER_HOST: 0.0.0.0
-      RENDER_TIMEOUT_MS: ${RENDER_TIMEOUT_MS:-15000}
+      RENDERER_PORT: "3000"   # keep in sync with the bot's RENDERER_URL and the healthcheck
-      MAX_OPTIONS: ${MAX_OPTIONS:-32}
+      RENDERER_MAX_CONCURRENT_RENDERS: ${RENDERER_MAX_CONCURRENT_RENDERS:-1}
-      MAX_OPTION_CHARS: ${MAX_OPTION_CHARS:-40}
+      RENDERER_RENDER_TIMEOUT_MS: ${RENDERER_RENDER_TIMEOUT_MS:-15000}
+      RENDERER_MAX_OPTIONS: ${RENDERER_MAX_OPTIONS:-32}
+      RENDERER_MAX_OPTION_CHARS: ${RENDERER_MAX_OPTION_CHARS:-40}
     expose:
       - "3000"
     healthcheck:

docs/deploy-coolify-selfhosted.md

@@ -60,8 +60,9 @@ compose network, so it needs no domain, publishes no port, and takes no auth
 token. Its API is unauthenticated, so never publish a port or attach a domain
 to it.
 
-Renderer tuning (`MAX_CONCURRENT_RENDERS`, `RENDER_TIMEOUT_MS`, `MAX_OPTIONS`,
+Renderer tuning (`RENDERER_MAX_CONCURRENT_RENDERS`,
-`MAX_OPTION_CHARS`) can be set in Coolify too; the defaults are in
+`RENDERER_RENDER_TIMEOUT_MS`, `RENDERER_MAX_OPTIONS`,
+`RENDERER_MAX_OPTION_CHARS`) can be set in Coolify too; the defaults are in
 [`renderer/docs/deployment.md`](../renderer/docs/deployment.md). Give the host
 1-2 GB of headroom for the renderer's Chrome.
 

renderer/.env.example

@@ -1,8 +1,8 @@
-PORT=3000
+RENDERER_PORT=3000
-HOST=0.0.0.0
+RENDERER_HOST=0.0.0.0
 NODE_ENV=production
-MAX_CONCURRENT_RENDERS=1
+RENDERER_MAX_CONCURRENT_RENDERS=1
 # Remotion browser timeout floor is 7000ms; lower values are raised to 7000.
-RENDER_TIMEOUT_MS=15000
+RENDERER_RENDER_TIMEOUT_MS=15000
-MAX_OPTIONS=32
+RENDERER_MAX_OPTIONS=32
-MAX_OPTION_CHARS=40
+RENDERER_MAX_OPTION_CHARS=40

renderer/README.md

@@ -67,7 +67,7 @@ result and its rarity — the service only draws it.
 
 Response is a silent H.264 `video/mp4` (Telegram plays it as an animation)
 with `X-Gacha-Rarity` and `X-Render-Duration-Ms` headers. All routes share the
-`MAX_CONCURRENT_RENDERS` slots. No game assets are used.
+`RENDERER_MAX_CONCURRENT_RENDERS` slots. No game assets are used.
 
 pack-cards animates on the browser clock, so the wish does not use Remotion
 compositions. `src/render/render-gacha.js` keeps one headless Chrome per
@@ -182,6 +182,6 @@ docker run --rm -p 3000:3000 miti99bot-renderer
 ```
 
 Recommended starting resources: 1-2 vCPU and 1-2 GB RAM, with
-`MAX_CONCURRENT_RENDERS=1`. `RENDER_TIMEOUT_MS` defaults to `15000` and is
+`RENDERER_MAX_CONCURRENT_RENDERS=1`. `RENDERER_RENDER_TIMEOUT_MS` defaults to
-raised to Remotion's `7000ms` browser timeout floor when configured lower.
+`15000` and is raised to Remotion's `7000ms` browser timeout floor when configured lower.
 The API has no authentication; publish its port only on a trusted network.

renderer/docs/deployment.md

@@ -28,15 +28,15 @@ Avoid for v1:
 
 ## Runtime
 
-Required env vars:
+Env vars, all optional (defaults shown):
 
 ```sh
-PORT=3000
+RENDERER_PORT=3000
-HOST=0.0.0.0
+RENDERER_HOST=0.0.0.0
-MAX_CONCURRENT_RENDERS=1
+RENDERER_MAX_CONCURRENT_RENDERS=1
-RENDER_TIMEOUT_MS=15000
+RENDERER_RENDER_TIMEOUT_MS=15000
-MAX_OPTIONS=32
+RENDERER_MAX_OPTIONS=32
-MAX_OPTION_CHARS=40
+RENDERER_MAX_OPTION_CHARS=40
 ```
 
 The root `compose.yml` forwards the tuning values with `${VAR:-default}`
@@ -50,7 +50,7 @@ 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.
 
-`RENDER_TIMEOUT_MS` is a total render timeout. Values below `7000` are raised to
+`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

renderer/src/config.js

@@ -30,14 +30,14 @@ const parsePositiveInt = (value, fallback) => {
  */
 export const loadConfig = (env = process.env) => {
   return {
-    host: env.HOST || '0.0.0.0',
+    host: env.RENDERER_HOST || '0.0.0.0',
-    port: parsePositiveInt(env.PORT, 3000),
+    port: parsePositiveInt(env.RENDERER_PORT, 3000),
-    maxConcurrentRenders: parsePositiveInt(env.MAX_CONCURRENT_RENDERS, 1),
+    maxConcurrentRenders: parsePositiveInt(env.RENDERER_MAX_CONCURRENT_RENDERS, 1),
     renderTimeoutMs: Math.max(
       minRenderTimeoutMs,
-      parsePositiveInt(env.RENDER_TIMEOUT_MS, 15000),
+      parsePositiveInt(env.RENDERER_RENDER_TIMEOUT_MS, 15000),
     ),
-    maxOptions: parsePositiveInt(env.MAX_OPTIONS, 32),
+    maxOptions: parsePositiveInt(env.RENDERER_MAX_OPTIONS, 32),
-    maxOptionChars: parsePositiveInt(env.MAX_OPTION_CHARS, 40),
+    maxOptionChars: parsePositiveInt(env.RENDERER_MAX_OPTION_CHARS, 40),
   };
 };

renderer/test/config.test.js

@@ -3,7 +3,7 @@ import {loadConfig, minRenderTimeoutMs} from '../src/config.js';
 
 describe('loadConfig', () => {
   test('keeps render timeout compatible with Remotion browser timeout limits', () => {
-    const config = loadConfig({RENDER_TIMEOUT_MS: '500'});
+    const config = loadConfig({RENDERER_RENDER_TIMEOUT_MS: '500'});
 
     expect(config.renderTimeoutMs).toBe(minRenderTimeoutMs);
   });