SQL Exporter

Features ​

  • Database-agnostic with broad driver support. The exporter supports a wide range of databases including PostgreSQL, MySQL, Microsoft SQL Server, and more.

  • Scrape-synchronous collection with optional caching. Metrics are collected on each Prometheus /metrics poll, which avoids the jitter, duplicate points and unscraped data of fixed-interval exporters. Minimal interval can be set per collector to serve cached results when scrapes are more frequent.

  • Staggered cache warmup. To prevent thundering herd effect on the database, the cache warmup is staggered on startup if configured.

  • Fully configuration-driven metrics. Metrics and the SQL queries behind them are defined in YAML. Queries are grouped into reusable collectors.

  • Flexible metric mapping. Each metric supports a flexible mapping from query results to Prometheus metrics to enrich the data with labels and values. The exporter can be configured to skip empty rows or rows with NULL label values.

  • Multi-target (jobs) mode. Single exporter instance can monitor many databases through a jobs list. Each job has named targets, shared labels and its own collector set.

  • Selective job scraping. The jobs[] query parameter lets Prometheus scrape particular jobs. This allows for more flexible scraping strategies.

  • Secret manager integration Connection string can reference a secret instead of holding static credentials for the following secret managers:

    • Kubernetes Secrets
    • AWS Secrets Manager
    • Google Cloud Secrets Manager
    • Hashicorp Vault

  • Secured HTTP/S endpoint. Use TLS certificates and basic authentication to secure the endpoint.

  • Request rate limiting. Limit the number of requests per second to avoid overload and protect the exporter from DoS attacks.

  • Live reload of configuration without restart. Reload the configuration file on SIGHUP signal or via HTTP endpoint without restarting the exporter.

  • Automatic metrics. Extra metrics for configured targets are collected automatically, including scrape duration, scrape errors, etc.

  • Builtin validation and config check.