Count hits, misses, fetches and transforms as Prometheus counters (closes #311)

The five counts the cache_stats table held are now Prometheus counters in
memory, in internal/imgcache/counters.go, on the default registry that
/metrics serves. The Increment methods add to them, so counting makes no
database call.

cache_stats, Cache.Stats, Service.Stats, CacheStats and the pending counts
are gone, with their tests. The background goroutine now writes only access
times, and its names say so: StartPendingAccessTimeWrites,
StopPendingAccessTimeWrites and pendingaccesstimes.go.

The counters count for the whole process, so the tests that check how much
they rose do not call t.Parallel. Two stop tests whose write under way was a
hit count now use an access time.

Model: opus-5-5

internal/imgcache/imgcache.go

@@ -44,23 +44,4 @@ type ImageCache interface {
 
 	// Warm pre-fetches and caches an image without returning it
 	Warm(ctx context.Context, req *imagerequest.Request) error
-
-	// Stats returns cache statistics
-	Stats(ctx context.Context) (*CacheStats, error)
-}
-
-// CacheStats contains cache statistics
-type CacheStats struct {
-	// TotalItems is the number of cached source images plus processed
-	// variants
-	TotalItems int64
-	// TotalSizeBytes is the total size of cached source images and
-	// processed variants
-	TotalSizeBytes int64
-	// HitCount is the number of cache hits
-	HitCount int64
-	// MissCount is the number of cache misses
-	MissCount int64
-	// HitRate is HitCount / (HitCount + MissCount)
-	HitRate float64
 }