NIP: Postgres connections high

Alert: NipPostgresConnectionsHigh · Severity: Warning

What this means

Active connections >80% of max_connections for 3+ minutes.

Fix

1. Grafana / Prometheus: pg_stat_activity_count, pg_settings_max_connections.
2. On server: docker exec inward_dev-db-1 psql -U $USER -d $DB -c "SELECT count(*), state FROM pg_stat_activity GROUP BY 2;"
3. Look for idle-in-transaction or connection leaks from web/celery.
4. Escalate to app team before raising max_connections blindly.

Escalate if

New transfers fail with "too many connections".