fix(swarm): adjust ports and configs (#1124)

Author: raisedadead

.gitignore

@@ -47,6 +47,7 @@ ansible/inventory/hosts
 
 # Secrets
 *.env
+*.env.*
 .envrc
 
 .kubeconfig.yaml

docker/swarm/stacks/socrates/.env.sample

@@ -5,14 +5,11 @@ DEPLOYMENT_VERSION=
 
 # -- Runtime
 NODE_ENV=production
-SERVER_URL=
 # ALLOWED_ORIGINS=*
 # LOG_LEVEL=info
 
 # -- Authentication
 API_KEY=
-DOCS_BASIC_AUTH_USER=
-DOCS_BASIC_AUTH_PASS=
 
 # -- Groq AI
 GROQ_API_KEY=

docker/swarm/stacks/socrates/stack-socrates.yml

@@ -1,6 +1,6 @@
 x-deploy-config: &deploy-config
   deploy:
-    replicas: 2
+    replicas: 3
     placement:
       max_replicas_per_node: 1
       constraints:
@@ -12,6 +12,8 @@ x-deploy-config: &deploy-config
       parallelism: 1
       delay: 10s
       failure_action: rollback
+      monitor: 30s
+      max_failure_ratio: 0.3
     resources:
       limits:
         cpus: ${CPU_LIMIT:-1}
@@ -22,77 +24,103 @@ x-deploy-config: &deploy-config
       condition: any
     labels:
       - org.freecodecamp.autoupdate=true
+  extra_hosts:
+    - "host.docker.internal:host-gateway"
+  healthcheck:
+    test:
+      [
+        "CMD-SHELL",
+        "wget --no-verbose --spider http://localhost:3001/health || exit 1",
+      ]
+    interval: 30s
+    timeout: 5s
+    start_period: 15s
+    start_interval: 5s
+    retries: 3
+  environment:
+    # Runtime
+    - NODE_ENV=${NODE_ENV:-production}
+    - PORT=3001
+    - ALLOWED_ORIGINS=${ALLOWED_ORIGINS:-*}
+    - LOG_LEVEL=${LOG_LEVEL:-info}
+    # Redis (local — each app connects to co-located Redis via host gateway)
+    - REDIS_URL=redis://host.docker.internal:6379
+    # Authentication
+    - API_KEY=${API_KEY}
+    # Groq AI
+    - GROQ_API_KEY=${GROQ_API_KEY}
+    - GROQ_MODEL=${GROQ_MODEL:-openai/gpt-oss-20b}
+    - GROQ_MODEL_HTML=${GROQ_MODEL_HTML:-openai/gpt-oss-20b}
+    - GROQ_MODEL_CSS=${GROQ_MODEL_CSS:-openai/gpt-oss-20b}
+    - GROQ_MODEL_JAVASCRIPT=${GROQ_MODEL_JAVASCRIPT:-openai/gpt-oss-120b}
+    - GROQ_MODEL_PYTHON=${GROQ_MODEL_PYTHON:-openai/gpt-oss-120b}
+    - GROQ_TIMEOUT_MS=${GROQ_TIMEOUT_MS:-30000}
+    - GROQ_MAX_RETRIES=${GROQ_MAX_RETRIES:-2}
+    - GROQ_BACKOFF_BASE_MS=${GROQ_BACKOFF_BASE_MS:-500}
+    - GROQ_MAX_TOKENS=${GROQ_MAX_TOKENS:-1024}
+    - GROQ_MAX_TOKENS_RETRY=${GROQ_MAX_TOKENS_RETRY:-2048}
+    - GROQ_EMPTY_RESPONSE_RETRIES=${GROQ_EMPTY_RESPONSE_RETRIES:-1}
+    # Circuit Breaker
+    - MODEL_CB_FAILURES=${MODEL_CB_FAILURES:-3}
+    - MODEL_CB_COOLDOWN_MS=${MODEL_CB_COOLDOWN_MS:-30000}
+    # Health Check
+    - ENABLE_EXTENDED_HEALTH=${ENABLE_EXTENDED_HEALTH:-false}
+    # Rate Limiting
+    - PER_USER_LIMIT=${PER_USER_LIMIT:-10}
+    - GLOBAL_LIMIT=${GLOBAL_LIMIT:-1000}
 
 x-port-config: &port-config
-  target: 3000
+  target: 3001
   protocol: tcp
   mode: host
 
 services:
-  svc-socrates:
+  svc-socrates-alpha:
     image: ${DOCKER_REGISTRY}/${DEPLOYMENT_ENV}/socrates:${DEPLOYMENT_VERSION}
     ports:
       - published: 4010
         <<: *port-config
     <<: *deploy-config
-    healthcheck:
-      test:
-        [
-          "CMD-SHELL",
-          "wget --no-verbose --spider http://localhost:3000/health || exit 1",
-        ]
-      interval: 30s
-      timeout: 5s
-      start_period: 15s
-      retries: 3
-    environment:
-      # Runtime
-      - NODE_ENV=${NODE_ENV:-production}
-      - PORT=3000
-      - SERVER_URL=${SERVER_URL}
-      - ALLOWED_ORIGINS=${ALLOWED_ORIGINS:-*}
-      - LOG_LEVEL=${LOG_LEVEL:-info}
-      # Redis (overlay network service discovery)
-      - REDIS_URL=redis://svc-redis:6379
-      # Authentication
-      - API_KEY=${API_KEY}
-      - DOCS_BASIC_AUTH_USER=${DOCS_BASIC_AUTH_USER}
-      - DOCS_BASIC_AUTH_PASS=${DOCS_BASIC_AUTH_PASS}
-      # Groq AI
-      - GROQ_API_KEY=${GROQ_API_KEY}
-      - GROQ_MODEL=${GROQ_MODEL:-openai/gpt-oss-20b}
-      - GROQ_MODEL_HTML=${GROQ_MODEL_HTML:-openai/gpt-oss-20b}
-      - GROQ_MODEL_CSS=${GROQ_MODEL_CSS:-openai/gpt-oss-20b}
-      - GROQ_MODEL_JAVASCRIPT=${GROQ_MODEL_JAVASCRIPT:-openai/gpt-oss-120b}
-      - GROQ_MODEL_PYTHON=${GROQ_MODEL_PYTHON:-openai/gpt-oss-120b}
-      - GROQ_TIMEOUT_MS=${GROQ_TIMEOUT_MS:-30000}
-      - GROQ_MAX_RETRIES=${GROQ_MAX_RETRIES:-2}
-      - GROQ_BACKOFF_BASE_MS=${GROQ_BACKOFF_BASE_MS:-500}
-      - GROQ_MAX_TOKENS=${GROQ_MAX_TOKENS:-1024}
-      - GROQ_MAX_TOKENS_RETRY=${GROQ_MAX_TOKENS_RETRY:-2048}
-      - GROQ_EMPTY_RESPONSE_RETRIES=${GROQ_EMPTY_RESPONSE_RETRIES:-1}
-      # Circuit Breaker
-      - MODEL_CB_FAILURES=${MODEL_CB_FAILURES:-3}
-      - MODEL_CB_COOLDOWN_MS=${MODEL_CB_COOLDOWN_MS:-30000}
-      # Health Check
-      - ENABLE_EXTENDED_HEALTH=${ENABLE_EXTENDED_HEALTH:-false}
-      # Rate Limiting
-      - PER_USER_LIMIT=${PER_USER_LIMIT:-10}
-      - GLOBAL_LIMIT=${GLOBAL_LIMIT:-1000}
+
+  svc-socrates-bravo:
+    image: ${DOCKER_REGISTRY}/${DEPLOYMENT_ENV}/socrates:${DEPLOYMENT_VERSION}
+    ports:
+      - published: 4011
+        <<: *port-config
+    <<: *deploy-config
 
   svc-redis:
     image: redis:7.4-alpine
-    command: ["redis-server", "--appendonly", "yes"]
+    ports:
+      - published: 6379
+        target: 6379
+        protocol: tcp
+        mode: host
+    command:
+      [
+        "redis-server",
+        "--appendonly",
+        "yes",
+        "--maxmemory",
+        "200mb",
+        "--maxmemory-policy",
+        "allkeys-lru",
+        "--bind",
+        "0.0.0.0",
+        "--protected-mode",
+        "no",
+      ]
     volumes:
       - redis_data:/data
     healthcheck:
       test: ["CMD", "redis-cli", "ping"]
       interval: 10s
       timeout: 5s
       start_period: 5s
+      start_interval: 5s
       retries: 3
     deploy:
-      replicas: 1
+      mode: global
       placement:
         constraints:
           - node.labels.socrates.enabled == true