Technical Details

Here are some technical details that might be interesting for those who are planning to use pgwatch for critical monitoring tasks or customize it in some way.

  • Dynamic management of monitored databases, metrics and their intervals - no need to restart/redeploy

    Config DB or YAML / SQL files are scanned every 2 minutes (by default, changeable via --servers-refresh-loop-seconds) and changes are applied dynamically. As common connectivity errors also handled, there should be no need to restart the gatherer "for fun". Please always report issues which require restarting.

  • There are some safety features built-in so that monitoring would not obstruct actual operation of databases

    • Up to 2 concurrent queries per monitored database (thus more per cluster) are allowed
    • Configurable statement timeouts per DB
    • SSL connections support for safe over-the-internet monitoring (use -e PW_WEBSSL=1 -e PW_GRAFANASSL=1 when launching Docker)
    • Optional authentication for the Web UI and Grafana (by default freely accessible)

  • Instance-level metrics caching

    To further reduce load on multi-DB instances, pgwatch can cache the output of metrics that are marked to gather only instance-level data. One such metric is for example "wal", and the metric attribute is "is_instance_level". Caching will be activated only for continuous source types, and to a default limit of up to 30 seconds (changeable via the --instance-level-cache-max-seconds param).