feat(memory): implement metrics and observability (#100)

Add comprehensive metrics collector for memory system with:
- Counter metrics: operations, retrievals, cache hits/misses, consolidations,
  episodes recorded, patterns/anomalies/insights detected
- Gauge metrics: item counts, memory size, cache size, procedure success rates,
  active sessions, pending consolidations
- Histogram metrics: working memory latency, retrieval latency, consolidation
  duration, embedding latency
- Prometheus format export
- Summary and cache stats helpers

31 tests covering all metric types, singleton pattern, and edge cases.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/app/services/memory/metrics/__init__.py

@@ -0,0 +1,18 @@
+# app/services/memory/metrics/__init__.py
+"""Memory Metrics module."""
+
+from .collector import (
+    MemoryMetrics,
+    get_memory_metrics,
+    record_memory_operation,
+    record_retrieval,
+    reset_memory_metrics,
+)
+
+__all__ = [
+    "MemoryMetrics",
+    "get_memory_metrics",
+    "record_memory_operation",
+    "record_retrieval",
+    "reset_memory_metrics",
+]