fix: Docker HEALTHCHECK reads port from config (#118) (#119)

* fix: Docker HEALTHCHECK reads port from config instead of hardcoded 8080 (#118)

HEALTHCHECK was hardcoded to localhost:8080. Users who changed
server.port in default.yaml (e.g. 8090) got an infinite restart loop
because the health check always failed.

Add docker-healthcheck.sh that parses port from config/default.yaml,
falling back to 8080 if unreadable.

* fix: remove unreachable exec in docker healthcheck script

---------

Co-authored-by: icebear0828 <icebear0828@users.noreply.github.com>

Dockerfile

@@ -37,10 +37,11 @@ EXPOSE 8080
 RUN mkdir -p /app/data
 
 COPY docker-entrypoint.sh /
-RUN chmod +x /docker-entrypoint.sh
+COPY docker-healthcheck.sh /
+RUN chmod +x /docker-entrypoint.sh /docker-healthcheck.sh
 
 HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
-  CMD curl -fs http://localhost:8080/health || exit 1
+  CMD /docker-healthcheck.sh
 
 ENTRYPOINT ["/docker-entrypoint.sh"]
 CMD ["node", "dist/index.js"]

docker-healthcheck.sh

@@ -0,0 +1,4 @@
+#!/bin/sh
+# Read server port from config, fallback to 8080
+PORT=$(grep -A5 '^server:' /app/config/default.yaml 2>/dev/null | grep 'port:' | head -1 | awk '{print $2}')
+curl -fs "http://localhost:${PORT:-8080}/health" || exit 1