Basira is not a generic APM with database plugins bolted on. Every collector, every metric, every view is designed for the engine it monitors — PostgreSQL and ClickHouse, today.
PostgreSQL
Query performance, table and index analytics, active query monitoring, wait events, checkpoints, I/O stats, and replication tracking.
ClickHouse
The only monitoring platform built specifically for ClickHouse OLAP workloads. Parts, merges, mutations, server metrics, and query analytics.
Across All Engines
Execution time breakdowns, call frequency, resource consumption, and historical trends for every normalized query.
Live WebSocket view of running queries, wait events, and lock dependencies. Auto-reconnection and configurable refresh rates.
Track replication lag, slot status, and replica health. Catch lag before it becomes user-facing.
Single Go binary deployed via Helm. Read-only access, minimal setup, ~20MB memory. Up and running in 5 minutes.
Set thresholds on key metrics, get notified via Slack, PagerDuty, Email, or Webhook.
Claude-driven SQL rewrites, missing index detection, and configuration tuning personalized to your workload.
From the Blog
Host-based and ingest-based pricing models explode at scale. Here's what's broken and how flat per-database pricing fixes it.
Index usage, join order, and common pitfalls in pg_stat_statements. Practical patterns for finding and fixing slow queries.
Parts, merges, and mutations: how the MergeTree engine actually works, and what to watch to keep OLAP workloads fast.