fix(database): correct metrics and improve error handling

- Fix metrics double-counting: track deltas for WaitCount/WaitDuration
  instead of adding cumulative values each tick
- Replace fmt.Printf with structured logging in pool monitor
- Add PoolOptions validation (MaxConns > 0, MinConns >= 0)
- Warn when DATABASE_URI overrides non-default PoolOptions
- Improve findAndParseConfig to report all tried files and errors
- Remove dead code in pgdb/config.go (unreachable host default)
- Fix errcheck lint issues for file.Close() calls
- Add context parameter to OpenDBMonitor() (breaking change)

database/metrics.go

@@ -3,10 +3,10 @@ package database
 import (
 	"context"
 	"database/sql"
-	"fmt"
 	"time"
 
 	"github.com/prometheus/client_golang/prometheus"
+	"go.ntppool.org/common/logger"
 )
 
 // DatabaseMetrics holds the Prometheus metrics for database connection pool monitoring
@@ -16,6 +16,10 @@ type DatabaseMetrics struct {
 	ConnectionsInUse        prometheus.Gauge
 	ConnectionsWaitCount    prometheus.Counter
 	ConnectionsWaitDuration prometheus.Histogram
+
+	// Track last values for delta calculation (cumulative stats from sql.DBStats)
+	lastWaitCount    int64
+	lastWaitDuration time.Duration
 }
 
 // NewDatabaseMetrics creates a new set of database metrics and registers them
@@ -67,26 +71,44 @@ func monitorConnectionPool(ctx context.Context, db *sql.DB, registerer prometheu
 	ticker := time.NewTicker(30 * time.Second)
 	defer ticker.Stop()
 
+	log := logger.FromContext(ctx)
+
 	for {
 		select {
 		case <-ctx.Done():
+			log.InfoContext(ctx, "database connection pool monitor stopped")
 			return
 		case <-ticker.C:
 			stats := db.Stats()
 
+			// Update gauge metrics (current state)
 			metrics.ConnectionsOpen.Set(float64(stats.OpenConnections))
 			metrics.ConnectionsIdle.Set(float64(stats.Idle))
 			metrics.ConnectionsInUse.Set(float64(stats.InUse))
-			metrics.ConnectionsWaitCount.Add(float64(stats.WaitCount))
 
-			if stats.WaitDuration > 0 {
-				metrics.ConnectionsWaitDuration.Observe(stats.WaitDuration.Seconds())
+			// Update counter with delta (WaitCount is cumulative in sql.DBStats)
+			waitCountDelta := stats.WaitCount - metrics.lastWaitCount
+			if waitCountDelta > 0 {
+				metrics.ConnectionsWaitCount.Add(float64(waitCountDelta))
+				metrics.lastWaitCount = stats.WaitCount
+			}
+
+			// Update histogram with delta (WaitDuration is cumulative in sql.DBStats)
+			waitDurationDelta := stats.WaitDuration - metrics.lastWaitDuration
+			if waitDurationDelta > 0 {
+				metrics.ConnectionsWaitDuration.Observe(waitDurationDelta.Seconds())
+				metrics.lastWaitDuration = stats.WaitDuration
 			}
 
 			// Log connection pool stats for high usage or waiting
-			if stats.OpenConnections > 20 || stats.WaitCount > 0 {
-				fmt.Printf("Connection pool stats: open=%d idle=%d in_use=%d wait_count=%d wait_duration=%s\n",
-					stats.OpenConnections, stats.Idle, stats.InUse, stats.WaitCount, stats.WaitDuration)
+			if stats.OpenConnections > 20 || waitCountDelta > 0 {
+				log.WarnContext(ctx, "high database connection usage",
+					"open", stats.OpenConnections,
+					"idle", stats.Idle,
+					"in_use", stats.InUse,
+					"wait_count", stats.WaitCount,
+					"wait_duration", stats.WaitDuration,
+				)
 			}
 		}
 	}