Neon Integration
Monitor serverless Postgres compute metrics, connection pooling, and storage branch health across your Neon projects. Get cold start detection and AI anomaly alerts without waking your suspended databases.
How It Works
Connect via Neon API
Generate a Neon API key in your project settings and add it to TigerOps. The API integration discovers all your projects, branches, and compute endpoints automatically — no agent required.
Enable Postgres Metrics Export
TigerOps connects to your Neon compute endpoint and collects pg_stat_activity, pg_stat_bgwriter, and connection pool metrics via the built-in statistics views. Read-only credentials are sufficient.
Configure Branch Health Alerts
Set up per-branch alerts for compute auto-suspension, storage divergence from the parent branch, and connection pool saturation. TigerOps tracks each branch independently for preview environment monitoring.
Map Compute Costs to Workloads
TigerOps tracks compute-active time, auto-suspend events, and storage growth per project. Correlate cost spikes with specific workloads, branches, or query patterns to optimize your Neon bill.
What You Get Out of the Box
Compute Endpoint Monitoring
Active compute duration, auto-suspend and resume events, cold start latency, and compute unit utilization per endpoint. Track which branches are burning compute unnecessarily.
Connection Pool Health
PgBouncer connection pool saturation, wait queue depth, idle connection counts, and transaction mode vs. session mode pooling efficiency for each Neon compute endpoint.
Storage Branch Health
Per-branch storage size, WAL accumulation rate, branch divergence from parent, and point-in-time restore history depth. Alert when branch storage costs exceed your budget.
Query Performance Tracking
pg_stat_statements integration for top-N slowest queries, query frequency, cache hit ratios, and lock wait times across all active branches in your Neon project.
Replication Lag Monitoring
For Neon projects using logical replication, TigerOps tracks replication slot lag, WAL sender throughput, and subscriber apply lag to keep your CDC pipelines healthy.
AI Cold Start Detection
TigerOps detects Neon compute cold start patterns and correlates them with application p99 latency spikes. Get recommendations for keep-alive strategies and optimal auto-suspend thresholds.
TigerOps Neon Integration Config
Connect TigerOps to your Neon projects via the API — no agent deployment needed.
# TigerOps Neon Integration Configuration
# API key from: https://console.neon.tech/app/settings/api-keys
integrations:
neon:
api_key_env: NEON_API_KEY
projects:
- project_id: "aged-snow-123456"
name: "production"
# Monitor only the main branch tightly
branch_policies:
- branch: "main"
alert_cold_start_ms: 500
alert_connection_pool_pct: 80
alert_storage_gb: 10
- branch_pattern: "preview/*"
alert_cold_start_ms: 2000
alert_connection_pool_pct: 95
- project_id: "cold-wind-789012"
name: "staging"
# Postgres stats collection (requires DB credentials)
postgres_metrics:
enabled: true
connection_string_env: NEON_PROD_CONN_STRING
collect_pg_stat_statements: true
top_queries: 20
poll_interval: 60s
# Auto-suspend monitoring
compute:
track_suspension_events: true
cold_start_alert_threshold_ms: 1000
exporters:
tigerops:
endpoint: "https://ingest.atatus.net/api/v1/write"
bearer_token: "${TIGEROPS_API_KEY}"Common Questions
Does TigerOps monitoring wake up my Neon compute?
TigerOps uses read-only Postgres connections that count as normal client connections. To avoid waking suspended computes unnecessarily, you can configure TigerOps to only poll active computes (those with recent activity) and skip polling suspended endpoints on a schedule.
Can TigerOps monitor Neon database branches separately?
Yes. TigerOps treats each Neon branch as an independent monitoring target. You can set different alert thresholds for your main branch, development branches, and ephemeral preview branches created by your CI/CD pipeline.
How does TigerOps handle Neon connection pooling metrics?
TigerOps collects PgBouncer statistics from the pgbouncer virtual database on your Neon endpoint. It tracks pool utilization, wait queue depth, and client connection rates in real time and alerts before your application hits connection exhaustion.
Can I track Neon compute costs per project in TigerOps?
Yes. TigerOps polls the Neon billing API to track compute-active hours and storage GB-hours per project. You can set budget alerts that fire when a project exceeds a monthly DBU spend threshold and attribute the spend to specific workloads.
Does TigerOps support Neon logical replication for CDC pipelines?
Yes. TigerOps monitors logical replication slot lag via pg_replication_slots and tracks WAL sender progress for each downstream subscriber. Alerts fire when replication falls behind your configured maximum lag threshold.
Stop Debugging Neon Cold Starts in Production
Cold start detection, branch health monitoring, and connection pool alerting. Connect in 5 minutes.