feat: implement Phase 3 scaling

- Replace skip-based pagination with cursor-based pagination (timestamp|_id cursors)
- Add Prometheus /metrics endpoint with request latency, fetch volume, and error counters
- Implement incremental fetch watermarking per source (watermarks collection in MongoDB)
- Add Graph change notification webhook endpoint (/api/webhooks/graph)
- Add correlation ID middleware for distributed tracing (x-request-id header)
- Update frontend to use cursor-based pagination with Prev/Next navigation
- Update tests for cursor pagination, metrics, webhooks, and watermark mocking

backend/metrics.py

@@ -0,0 +1,43 @@
+
+from prometheus_client import Counter, Histogram, generate_latest
+
+REQUEST_DURATION = Histogram(
+    "aoc_request_duration_seconds",
+    "HTTP request duration",
+    ["method", "path", "status"],
+)
+EVENTS_FETCHED = Counter(
+    "aoc_events_fetched_total",
+    "Number of audit events fetched per source",
+    ["source"],
+)
+FETCH_ERRORS = Counter(
+    "aoc_fetch_errors_total",
+    "Number of fetch errors per source",
+    ["source"],
+)
+FETCH_DURATION = Histogram(
+    "aoc_fetch_duration_seconds",
+    "Duration of fetch jobs per source",
+    ["source"],
+)
+
+
+def observe_request(method: str, path: str, status: int, duration: float):
+    REQUEST_DURATION.labels(method=method, path=path, status=str(status)).observe(duration)
+
+
+def track_fetch(source: str, count: int):
+    EVENTS_FETCHED.labels(source=source).inc(count)
+
+
+def track_fetch_error(source: str):
+    FETCH_ERRORS.labels(source=source).inc()
+
+
+def track_fetch_duration(source: str, duration: float):
+    FETCH_DURATION.labels(source=source).observe(duration)
+
+
+def prometheus_metrics():
+    return generate_latest()