feat: implement observability and metrics (Task 22)

- Add metrics package tracking cache hit-rate, API quota, circuit breaker trips, search time
- Integrate metrics with cache layer, yandex client, and API handlers
- Add /metrics HTTP endpoint for Prometheus-compatible metrics exposure
- Write tests for metrics functionality across cache, yandex, and API handlers
- Update test files to support new metrics infrastructure

internal/routing/search_cache.go

@@ -5,20 +5,23 @@ import (
 	"fmt"
 
 	"trip-planner/internal/cache"
+	"trip-planner/internal/metrics"
 	"trip-planner/internal/yandex"
 )
 
 // SearchCacheService handles caching and on-demand Yandex /search calls.
 type SearchCacheService struct {
-	cache  *cache.CacheAside
-	yclient *yandex.Client
+	cache    *cache.CacheAside
+	yclient  *yandex.Client
+	metrics  *metrics.Metrics
 }
 
 // NewSearchCacheService creates a new search cache service.
-func NewSearchCacheService(cacheStore cache.Cache, yclient *yandex.Client) *SearchCacheService {
+func NewSearchCacheService(cacheStore cache.Cache, yclient *yandex.Client, m *metrics.Metrics) *SearchCacheService {
 	return &SearchCacheService{
-		cache:   cache.NewCacheAside(cacheStore),
+		cache:   cache.NewCacheAside(cacheStore, m),
 		yclient: yclient,
+		metrics: m,
 	}
 }