switched from healthchecks to status checks

calendar/Dockerfile

@@ -24,5 +24,5 @@ ENV     GIT_COMMIT_SHA=${GIT_COMMIT_SHA}
 EXPOSE  8083
 
 ENTRYPOINT ["tini", "--"]
-HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8083/api/health || exit 1
+HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8083/api/calendar/status || exit 1
 CMD     /bin/navigatum-calendar

calendar/src/main.rs

@@ -7,11 +7,15 @@ use actix_web::{get, middleware, web, App, HttpResponse, HttpServer};
 
 const MAX_JSON_PAYLOAD: usize = 1024 * 1024; // 1 MB
 
-#[get("/api/calendar/health")]
-async fn health_handler() -> HttpResponse {
+#[get("/api/calendar/status")]
+async fn health_status_handler() -> HttpResponse {
+    let github_link = match std::env::var("GIT_COMMIT_SHA") {
+        Ok(hash) => format!("https://github.com/TUM-Dev/navigatum/tree/{hash}"),
+        Err(_) => "unknown commit hash, probably running in development".to_string(),
+    };
     HttpResponse::Ok()
         .content_type("text/plain")
-        .body("healthy")
+        .body(format!("healthy\nsource_code: {github_link}"))
 }
 
 #[actix_web::main]
@@ -33,7 +37,7 @@ async fn main() -> std::io::Result<()> {
             .wrap(middleware::Logger::default())
             .wrap(middleware::Compress::default())
             .app_data(web::JsonConfig::default().limit(MAX_JSON_PAYLOAD))
-            .service(health_handler)
+            .service(health_status_handler)
     })
     .bind(std::env::var("BIND_ADDRESS").unwrap_or_else(|_| "0.0.0.0:8060".to_string()))?
     .run()

deployment/k3s/templates/deployments/calendar-deployment.yaml

@@ -57,13 +57,13 @@ spec:
               memory: 500Mi
           livenessProbe:
             httpGet:
-              path: /api/calendar/health
+              path: /api/calendar/status
               port: calendar
             failureThreshold: 2
             periodSeconds: 10
           startupProbe:
             httpGet:
-              path: /api/calendar/health
+              path: /api/calendar/status
               port: calendar
             failureThreshold: 60
             periodSeconds: 1

deployment/k3s/templates/deployments/feedback-deployment.yaml

@@ -46,13 +46,13 @@ spec:
               memory: 40Mi
           livenessProbe:
             httpGet:
-              path: /api/feedback/health
+              path: /api/feedback/status
               port: feedback
             failureThreshold: 2
             periodSeconds: 1
           startupProbe:
             httpGet:
-              path: /api/feedback/health
+              path: /api/feedback/status
               port: feedback
             failureThreshold: 6
             periodSeconds: 1

deployment/k3s/templates/deployments/server-deployment.yaml

@@ -70,13 +70,13 @@ spec:
               memory: 700Mi
           livenessProbe:
             httpGet:
-              path: /api/health
+              path: /api/status
               port: api
             failureThreshold: 1
             periodSeconds: 1
           startupProbe:
             httpGet:
-              path: /api/health
+              path: /api/status
               port: api
             failureThreshold: 60
             periodSeconds: 1

feedback/Dockerfile

@@ -23,5 +23,5 @@ ENV     GIT_COMMIT_SHA=${GIT_COMMIT_SHA}
 EXPOSE  8070
 
 ENTRYPOINT ["tini", "--"]
-HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8070/api/feedback/health || exit 1
+HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8070/api/feedback/status || exit 1
 CMD     /bin/navigatum-feedback

feedback/src/main.rs

@@ -19,11 +19,15 @@ pub struct Opt {
     jwt_key: Option<String>,
 }
 
-#[get("/api/feedback/health")]
-async fn health_handler() -> HttpResponse {
+#[get("/api/feedback/status")]
+async fn health_status_handler() -> HttpResponse {
+    let github_link = match std::env::var("GIT_COMMIT_SHA") {
+        Ok(hash) => format!("https://github.com/TUM-Dev/navigatum/tree/{hash}"),
+        Err(_) => "unknown commit hash, probably running in development".to_string(),
+    };
     HttpResponse::Ok()
         .content_type("text/plain")
-        .body("healthy")
+        .body(format!("healthy\nsource_code: {github_link}"))
 }
 
 #[tokio::main]
@@ -48,10 +52,10 @@ async fn main() -> std::io::Result<()> {
 
         App::new()
             .wrap(cors)
-            .wrap(middleware::Logger::default().exclude("/api/feedback/health"))
+            .wrap(middleware::Logger::default().exclude("/api/feedback/status"))
             .wrap(middleware::Compress::default())
             .app_data(web::JsonConfig::default().limit(MAX_JSON_PAYLOAD))
-            .service(health_handler)
+            .service(health_status_handler)
             .service(
                 web::scope("/api/feedback")
                     .configure(core::configure)

openapi.yaml

@@ -937,7 +937,7 @@ paths:
                   - Not found
       tags:
         - migration
-  /api/health:
+  /api/status:
     get:
       operationId: api-health
       summary: API healthcheck
@@ -950,14 +950,15 @@ paths:
             text/plain:
               schema:
                 type: string
-                enum:
-                  - healthy
+                example: |
+                  healthy
+                  source code: https://github.com/TUM-Dev/navigatum/tree/8a0fb71819ac88c8af35683cfb46291f0d0c9b0a
         '503':
           description: Service Unavailable
           content: {}
       tags:
         - health
-  /api/feedback/health:
+  /api/feedback/status:
     get:
       operationId: feedback-health
       summary: feedback-API healthcheck
@@ -970,8 +971,30 @@ paths:
             text/plain:
               schema:
                 type: string
-                enum:
-                  - healthy
+                example: |
+                  healthy
+                  source code: https://github.com/TUM-Dev/navigatum/tree/8a0fb71819ac88c8af35683cfb46291f0d0c9b0a
+        '503':
+          description: Service Unavailable
+          content: {}
+      tags:
+        - health
+  /api/calendar/status:
+    get:
+      operationId: calendar-health
+      summary: calendar-API healthcheck
+      description: |
+        If this endpoint does not return 200, the API is experiencing a catastrophic outage. Should never happen.
+      responses:
+        '200':
+          description: Ok
+          content:
+            text/plain:
+              schema:
+                type: string
+                example: |
+                  healthy
+                  source code: https://github.com/TUM-Dev/navigatum/tree/8a0fb71819ac88c8af35683cfb46291f0d0c9b0a
         '503':
           description: Service Unavailable
           content: {}

server/Dockerfile.server

@@ -26,5 +26,5 @@ ENV     GIT_COMMIT_SHA=${GIT_COMMIT_SHA}
 EXPOSE  8080
 
 ENTRYPOINT ["tini", "--"]
-HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8080/api/health || exit 1
+HEALTHCHECK --start-period=20m  --timeout=10s CMD curl --fail localhost:8080/api/status || exit 1
 CMD     /bin/navigatum-server

server/src/main.rs

@@ -9,7 +9,7 @@ mod utils;
 
 const MAX_JSON_PAYLOAD: usize = 1024 * 1024; // 1 MB
 
-#[get("/api/health")]
+#[get("/api/status")]
 async fn health_handler() -> HttpResponse {
     HttpResponse::Ok()
         .content_type("text/plain")
@@ -29,7 +29,7 @@ async fn main() -> std::io::Result<()> {
 
         App::new()
             .wrap(cors)
-            .wrap(middleware::Logger::default().exclude("/api/health"))
+            .wrap(middleware::Logger::default().exclude("/api/status"))
             .wrap(middleware::Compress::default())
             .app_data(web::JsonConfig::default().limit(MAX_JSON_PAYLOAD))
             .service(health_handler)