Cluster Status
Healthy
6 Nodes · 1 primary + 2 sync standby + 3 async replica
No failovers · 30d
Repl lag 0.4 ms
Database Console & Query Monitor
Live cluster telemetry, heavy-query watchlist and NVMe storage pressure for production-analytics-v2 on fi-central-main.
Streaming
Query Response Time
4.2ms
avg over 5 min · p95 11.8 ms · 18 420 q/s
Target budget 12 ms · headroom 65%
Traffic Today
Ingress
62.4 GB
Egress Today
148.6 GB
Daily quota 200 GB · resets 00:00 UTC
Pooled Connections
342
clients · 48 server conns · cache hit 99.2%
1 452 txn/s
0 aborted
Open pool & VPC detail
Query Latency — last 60 seconds
pg_stat_statements rolling mean, sampled every 2 s
mean
p95
-60s-45s-30s-15snow
Node Topology
fi-central-main · 3 availability zones
NVMe Storage Subsystem Load
Per-node aggregated device pressure, 5 s sample interval
iostat -x 5
queue depth 4.1
NVMe Read
1.84GB/s
IOPS 412 900
latency 0.09 ms
NVMe Write
0.94GB/s
IOPS 186 400
WAL fsync p99 0.21 ms
| Volume | Read | Write | %util | Wear |
|---|
Devices are local NVMe (XFS, /var/lib/pgsql/16/data), TRIM weekly, 3× replica on object storage.
Running Heavy Queries
Watchlist threshold: memory > 512 MB or runtime > 60 s · 6 active
| Query ID | Database | User | Memory | Execution Time | State / Node | Action |
|---|
Resolution: pg_stat_activity +
pg_stat_statements · terminate issues
pg_terminate_backend(pid, 15000)
Auto-kill policy: runtime > 900 s on replica role only
Checkpoint & WAL
- Checkpoints (timed / requested)
- 412 / 3
- WAL generated today
- 86.4 GB
- WAL archive lag
- 0.4 s
- Buffers per checkpoint
- 1 024 × 8 kB
- Deadlocks (24h)
- 0