Fix nginx crash-loop on a fully fresh docker compose up

nginx resolves a plain `proxy_pass http://server:8080;` hostname once at
config-load time. On a cold `docker compose up` that creates every
container at once, the `server` container isn't always registered in
Docker's embedded DNS yet by the time the client's nginx starts — nginx
then fails immediately ("host not found in upstream") and doesn't retry,
so it just crash-loops.

Route the target through a variable instead (`set $upstream_server ...;
proxy_pass $upstream_server;`) with an explicit `resolver`, which makes
nginx resolve the hostname lazily per-request via Docker's DNS rather than
once at startup. Also add `depends_on: [server]` on the client service so
it at least doesn't start before the server container exists at all.

Verified with repeated `docker compose down && docker compose up` cycles
locally — this only reliably reproduced starting every container from
nothing simultaneously, which prior local testing hadn't actually done
(the server container had usually stayed running across rebuilds).

docker-compose.yml

@@ -37,6 +37,11 @@ services:
     build:
       context: .
       dockerfile: src/Client/Dockerfile
+    # Doesn't fix the DNS race on its own (nginx.conf's lazy resolver does
+    # that) — just avoids nginx crashing into a restart loop for no reason
+    # by not starting client before `server` exists at all.
+    depends_on:
+      - server
     # Loopback-only in production, where a host-level nginx (TLS + the real
     # domain) is the actual public entry point and proxies here — see
     # docs/DEPLOY.md. For local `docker compose up`, still reachable at

src/Client/nginx.conf

@@ -2,11 +2,23 @@ server {
     listen 80;
     root /usr/share/nginx/html;
 
-    # No trailing slash on proxy_pass here — the full matched URI (including
+    # `server` is resolved as a variable (not a plain hostname in proxy_pass)
-    # the /api/ prefix) is forwarded unchanged, matching Giraffe's route
+    # so nginx uses Docker's embedded DNS resolver *lazily, per request*
-    # definitions in Server/Routes.fs, which all start with /api/.
+    # instead of resolving it once at config-load time. A static
+    # `proxy_pass http://server:8080;` fails permanently at nginx startup
+    # ("host not found in upstream") if the `server` container isn't already
+    # registered in Docker's DNS yet — a real race on a fully fresh
+    # `docker compose up` that starts every container at once, since nginx
+    # doesn't retry once it's already crashed.
+    resolver 127.0.0.11 valid=10s;
+
+    # No trailing slash on the proxy_pass target — the full matched URI
+    # (including the /api/ prefix) is forwarded unchanged, matching
+    # Giraffe's route definitions in Server/Routes.fs, which all start
+    # with /api/.
     location /api/ {
-        proxy_pass http://server:8080;
+        set $upstream_server http://server:8080;
+        proxy_pass $upstream_server;
         proxy_set_header Host $host;
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;