Database Monitoring
Database monitoring surfaces slow queries, lock contention, connection saturation, and replication lag — and ties each back to the application transaction that triggered it.
Supported engines#
PostgreSQL, MySQL/MariaDB, Oracle, SQL Server, MongoDB, and Redis. Enable collection by adding a read-only monitoring user and pointing the agent at the instance.
applicare-agent config set db.postgres.dsn="host=db1 user=arc_ro dbname=app sslmode=require"Query performance#
Applicare captures normalised query fingerprints with execution counts, latency percentiles, rows examined, and plan changes over time. The Top queries view ranks statements by total time so you fix the query that actually moves the needle.
Correlation: every captured query links to the traces that ran it, so you can jump from a slow endpoint straight to the offending statement and back.
Health & saturation#
Track connections vs. max, buffer/cache hit ratios, replication lag, deadlocks, and long-running transactions. Baseline-driven alerts fire when a metric deviates from its learned norm rather than a static guess. See Database solutions for engine-specific playbooks.