Skip to main content

Development

The scraper is a Go program that connects to Oracle databases on a schedule and writes collected samples into PostgreSQL.

The runtime flow is:

  • Parse --config.file and expand environment variables.
  • Load database, metric, PostgreSQL output, logging, and web configuration.
  • Load any optional files listed under metrics.definitions.
  • Open Oracle connection pools for each configured source database.
  • Open the PostgreSQL connection pool and run schema migrations when output.postgresql.autoMigrate is enabled.
  • Start the lightweight web listener for /healthz and /.
  • Start scheduled native collectors for SQL, cached cursor plans, sessions, blocking, DAH, and alert logs, plus any configured additional metrics.

Generic TOML/YAML metrics are written to oracle_metric_samples. Performance collectors write to dedicated tables such as oracle_sql_samples, oracle_sql_texts, oracle_sql_plans, oracle_session_samples, oracle_blocking_session_samples, and oracle_database_activity_samples.

Grafana dashboards read PostgreSQL directly. The scraper does not expose Oracle metrics on /metrics.

Container Build​

To build a container image, run:

make docker

For ARM:

make docker-arm

Binary Build​

The default build uses the godror driver:

go build -o oracledb_performance_scraper ./

Build with the pure-Go go-ora driver:

go build -tags goora -o oracledb_performance_scraper ./

The Makefile release build writes binaries and archives under dist.