split docker-compose into core and observability stacks; add ntfy, alertmanager, blackbox exporter, and otel-collector services; configure prometheus alerting pipeline

prometheus/alert.rules.yml

@@ -0,0 +1,11 @@
+groups:
+  - name: uptime_alerts
+    rules:
+      - alert: ServiceDown
+        expr: probe_success{job="blackbox_http"} == 0
+        for: 1m
+        labels:
+          severity: critical
+        annotations:
+          summary: "{{ $labels.instance }} est down"
+          description: "Le service {{ $labels.instance }} ne répond plus depuis plus d'1 minute."