Telemetry, Golden Signals & Alerting Thresholds
Database Monitoring & Alert Thresholds
Eliminate alert fatigue by setting actionable SLA-driven threshold rules for Prometheus, Datadog, and CloudWatch.
Filter by Engine:
Category:
| Kritik Metrik & Kategori | Motor | Uyarı (Warning) | Kritik (Pager Alert) | Önem | Açıklama & Risk | Alınacak Aksiyon |
|---|---|---|---|---|---|---|
Replication Lag (Byte / Saniye Gecikmesi) Replication | PostgreSQL | > 50 MB veya > 10 sn | > 500 MB veya > 60 sn | 🔴 Critical | WAL byte difference or replay time lag between Primary and Standby replica nodes. | Check replica CPU/IOPS utilization. Terminate blocking analytic queries on standby or optimize network throughput. |
Connection Pool Saturation (Bağlantı Doluluğu) Connections | PostgreSQL | max_connections %75 | max_connections %90 | 🔴 Critical | Ratio of active and idle client connections relative to configured max_connections. | Tune PgBouncer pool sizing. Enforce idle_in_transaction_session_timeout to terminate hanging web worker sessions. |
Disk Free Space (Boş Depolama Alanı) Storage | PostgreSQL | < %20 boş alan | < %10 boş alan | 🔴 Critical | Remaining percentage of available storage on database data and WAL mount partitions. | Prune archived WAL files, drop stale temp tables, and expand block volume capacity immediately. |
Transaction ID Wraparound Age (datfrozenxid) Maintenance | PostgreSQL | > 1.0 Milyar XID | > 1.5 Milyar XID | 🔴 Critical | Current transaction ID consumption relative to the 2-billion wraparound safety threshold. | Trigger aggressive VACUUM FREEZE on oldest tables and increase autovacuum_cost_limit throttle. |
Buffer Cache Hit Ratio (Bellek İsabet Oranı) Performance | PostgreSQL | < %98 | < %95 | 🟠 High | Percentage of page read requests served directly from shared_buffers without triggering disk I/O. | Identify queries with highest disk blks_read via pg_stat_statements; add covering indexes or increase shared_buffers. |
CPU Utilization (İşlemci Yükü) Performance | PostgreSQL | > %75 (5 dk ortalama) | > %90 (3 dk sürekli) | 🟠 High | Overall CPU utilization and system load average on the primary database host. | Inspect pg_stat_activity for runaway sequential scans or complex joins consuming all CPU cores. |
Disk I/O Latency (Read / Write Gecikmesi) Storage | PostgreSQL | > 10 ms | > 25 ms | 🟠 High | Average read/write response latency in milliseconds reported by the storage layer. | Throttle checkpoint and autovacuum write rates; provision higher IOPS / throughput bandwidth on cloud storage. |
Deadlock Rate (Ölümcül Kilitlenme Sıklığı) Errors | PostgreSQL | > 1 / 5 dk | > 5 / 1 dk | 🔴 Critical | Frequency of conflicting transactional locks detected and rolled back by the deadlock detector. | Enforce deterministic lock acquisition ordering across application transactions and analyze deadlock log traces. |
Temporary File Generation Rate (Geçici Disk Dosyaları) Performance | PostgreSQL | > 50 MB / sorgu | > 500 MB / sorgu | 🟡 Medium | Volume of temporary spill files created on disk when queries exceed work_mem limits. | Increase work_mem for specific analytic sessions or create B-tree indexes to enable index-based sorting. |
Long Running Transactions / Queries (Uzun Süren Sorgular) Performance | PostgreSQL | > 120 saniye | > 600 saniye | 🟠 High | Number of active transactions running continuously beyond acceptable operational duration thresholds. | Enforce statement_timeout limits and cancel runaway analytical queries blocking transactional OLTP workloads. |
Seconds Behind Source (MySQL Replikasyon Gecikmesi) Replication | MySQL | > 10 saniye | > 60 saniye | 🔴 Critical | Seconds Behind Master lag reported by MySQL replica SQL_THREAD during binlog execution. | Enable Multi-Threaded Replication (replica_parallel_workers) and tune replica I/O buffering. |