squash

Author: pwnage101

common.yml

@@ -28,12 +28,19 @@ services:
           done
     # Configure a healthcheck to allow `make dev.up.<servce>` to give accurate feedback about
     # whether the service actually started successfully and is serving traffic.
-    # Set HEALTHCHECK_TARGET to the full URL curl should hit, e.g. "http://localhost:18000/heartbeat"
+    #
+    # Set HEALTHCHECK_TARGET to the full URL curl should hit, e.g.:
+    #
+    #   environment:
+    #     HEALTHCHECK_TARGET: "http://localhost:18150/health/"
+    #
+    # Timings are sized so the worst case from container start to "give up"
+    # (marked unhealthy) is ~5 minutes: start_period + retries*interval = 300s = 5min.
     healthcheck:
       test: ["CMD-SHELL", "curl -fsS $$HEALTHCHECK_TARGET || exit 1"]
       interval: 10s
       timeout: 10s
-      retries: 20
+      retries: 6
       start_period: 240s
       start_interval: 5s  # poll fast during grace period so a healthy boot flips status quickly
 
@@ -68,11 +75,13 @@ services:
     image: node:18
     environment:
       - NODE_ENV=development
-    # More generous than backend-app because a cold `npm ci` can take
-    # several minutes.
+    # MFEs take longer to startup: a cold `npm ci` + `npm start` is empirically
+    # observed to reach healthy in about 5 minutes, so we shouldn't give up
+    # until at least ~10 minutes. start_period + retries*interval = 600s = 10min.
     healthcheck:
       test: ["CMD-SHELL", "curl -fsS $$HEALTHCHECK_TARGET || exit 1"]
       interval: 10s
       timeout: 10s
-      retries: 30
-      start_period: 300s
+      retries: 6
+      start_period: 540s
+      start_interval: 5s  # poll fast during grace period so a healthy boot flips status quickly