attempt to minimize disk wear from Redis

fix peanut healthcheck

Author: romkey

apps/authentik/README.md

@@ -4,6 +4,10 @@ Identity Provider and SSO solution
 
 ## Configuration
 
+### Redis (`authentik-redis`)
+
+Redis runs with **no AOF** and **no RDB** (`save ""`) to limit SSD wear. Cache and sessions are **volatile** across unclean restarts—users may need to **sign in again**. See [redis-persistence-hackstack.md](../../docs/redis-persistence-hackstack.md).
+
 ### Environment Variables
 
 Copy `.env.example` to `.env` and configure:

apps/authentik/docker-compose.yml

@@ -3,7 +3,15 @@ services:
     image: docker.io/library/redis:${REDIS_IMAGE_VERSION:-alpine}
     container_name: authentik-redis
     hostname: authentik-redis
-    command: --save 60 1 --loglevel warning
+    # No AOF, no RDB: cache/sessions only; users re-login after unclean restart (minimal SSD wear).
+    command:
+      - redis-server
+      - --save
+      - ""
+      - --appendonly
+      - "no"
+      - --loglevel
+      - warning
     restart: unless-stopped
     healthcheck:
       test: ["CMD-SHELL", "redis-cli ping | grep PONG"]

apps/event-manager/README.md

@@ -11,7 +11,7 @@ Image: [romkey/pdxhackerspace-eventmanager](https://github.com/romkey/pdxhackers
 |-----------|------|
 | `event-manager` | Rails web server (port 3000, behind reverse proxy) |
 | `event-manager-sidekiq` | Sidekiq background job worker |
-| `event-manager-redis` | Private Redis instance for job queuing (not shared) |
+| `event-manager-redis` | Private Redis for Sidekiq (no AOF; `save 900 1` only—see [redis-persistence-hackstack.md](../../docs/redis-persistence-hackstack.md)) |
 
 ## Network dependencies
 

apps/event-manager/docker-compose.yml

@@ -30,6 +30,14 @@ services:
     container_name: event-manager-redis
     hostname: event-manager-redis
     restart: unless-stopped
+    # No AOF; RDB at most every 15m if keys changed (Sidekiq queue may be replay-lost on hard crash).
+    command:
+      - redis-server
+      - --appendonly
+      - "no"
+      - --save
+      - "900"
+      - "1"
     volumes:
       - ../../lib/event-manager/redis:/data
     networks:

apps/glitchtip/README.md

@@ -10,7 +10,7 @@ PostgreSQL and Redis — no ClickHouse, Kafka, or Snuba required.
 |---|---|---|
 | `glitchtip` | `glitchtip/glitchtip` | Django web app + API |
 | `glitchtip-worker` | `glitchtip/glitchtip` | Celery worker + beat scheduler |
-| `glitchtip-redis` | `redis` | Celery broker and cache |
+| `glitchtip-redis` | `redis` | Celery broker and cache (no AOF; `save 900 1` only—see [redis-persistence-hackstack.md](../../docs/redis-persistence-hackstack.md)) |
 | `glitchtip-migrate` | `glitchtip/glitchtip` | One-shot database migration runner |
 
 ## Networks

apps/glitchtip/docker-compose.yml

@@ -4,6 +4,14 @@ services:
     container_name: glitchtip-redis
     hostname: glitchtip-redis
     restart: unless-stopped
+    # No AOF; RDB at most every 15m if keys changed (Celery queue may be replay-lost on hard crash).
+    command:
+      - redis-server
+      - --appendonly
+      - "no"
+      - --save
+      - "900"
+      - "1"
     volumes:
       - ../../lib/glitchtip/redis:/data
     networks:

apps/member-manager/README.md

@@ -11,7 +11,7 @@ Image: [romkey/pdxhackerspace-member-manager](https://github.com/romkey/pdxhacke
 |-----------|------|
 | `member-manager` | Rails web server (port 3000, behind reverse proxy) |
 | `member-manager-sidekiq` | Sidekiq background job worker |
-| `member-manager-redis` | Private Redis instance for job queuing (not shared) |
+| `member-manager-redis` | Private Redis for Sidekiq (no AOF; `save 900 1` only—see [redis-persistence-hackstack.md](../../docs/redis-persistence-hackstack.md)) |
 
 ## Network dependencies
 

apps/member-manager/docker-compose.yml

@@ -4,6 +4,14 @@ services:
     container_name: member-manager-redis
     hostname: member-manager-redis
     restart: unless-stopped
+    # No AOF; RDB at most every 15m if keys changed (Sidekiq queue may be replay-lost on hard crash).
+    command:
+      - redis-server
+      - --appendonly
+      - "no"
+      - --save
+      - "900"
+      - "1"
     volumes:
       - ../../lib/member-manager/redis:/data
     networks:

apps/peanut/README.md

@@ -17,6 +17,10 @@ cp .env.example .env
 
 Configuration files are stored in the `config/` directory.
 
+## Healthcheck
+
+The image defines a working check: `node healthcheck.mjs` hits `http://127.0.0.1:$WEB_PORT/api/ping`. Do not use `wget`/`curl` probes—the runtime image is minimal and may not include them.
+
 ## Usage
 
 ### Starting the service

apps/peanut/docker-compose.yml

@@ -12,12 +12,7 @@ services:
       - proxy
     env_file:
       - .env
-    healthcheck:
-      test: ["CMD-SHELL", "wget -q --spider http://127.0.0.1:8080/ || exit 1"]
-      interval: 30s
-      timeout: 10s
-      retries: 3
-      start_period: 45s
+    # Use image HEALTHCHECK: `node healthcheck.mjs` → GET /api/ping (no wget in image; hard-coded :8080 wget fails if WEB_PORT differs).
 
 networks:
   proxy: