oracle/oracle-db-appdev-monitoring 2.3.0

on GitHub

2.3.0

• Add per-database scrape duration metrics to complement the existing exporter-wide scrape duration metric.
• Add per-database scrape error counters while preserving the existing global scrape error counter.
• Normalize loaded metric identifiers to lowercase so mixed-case custom metric names, descriptions, labels, histogram fields, and fieldtoappend values are handled correctly.
• Alert log records now include their source database name in the exported JSON output.
• Add the log.perDatabaseFiles configuration option to write alert logs to per-database files such as alert-db2.log. This is recommended when scraping multiple databases from a single exporter.
• Document that TLS, basic authentication, and other Prometheus web server security settings should be configured through web.configFile using Prometheus Exporter Toolkit configuration.
• Start the exporter web server before initializing database connections and warming connection pools, so unavailable databases do not block the metrics endpoint from coming up.
• Improve database ping handling so transient connection errors enter backoff and closed connections are rebuilt and warmed before reuse.
• Fix alert log export when nullable database fields such as execution_context_id are returned as NULL.
• Fix metrics caching where metrics scrape failures could improperly invalidate entries in the metrics cache.
• Fix a metrics file-loading issue an invalid metrics file failed to fall on the packaged default metrics.
• Fix a username/password encoding bug when using the go-ora driver.
• Update third-party dependencies.
• Update Go runtime to 1.26.2.

Thank you to the following users for their suggestions and contributions:

• @wes-pro
• @MansuyDavid
• baranchn
• @kirxcore
• finnure

Full Changelog: 2.2.2...2.3.0