Architecture

PATROL for PostgreSQL enables you to monitor local or remote PostgreSQL database instances.

Local monitoring

Remote monitoring

Data collection details

PATROL for PostgreSQL collects the data

In local monitoring, PATROL for PostgreSQL starts a Java process. This Java process runs queries on the PostgreSQL database to generate a list of instances and databases.
For remote monitoring, the Java process identifies where the remote server is running. Based on the operating system, the Java process creates a remote client connection. For UNIX, the remote connection is an SSH channel. For Windows, the remote client connection is an SSH or WMI channel. The same process logic applies as local monitoring.
In the first inventory poll, the Java process discovers the PostgreSQL system (databases and table spaces) and creates the discovery tree for PATROL.
After the discovery is complete, the Java collector starts collecting availability for the PostgreSQL instances, databases and partitions. The default availability poll cycle is 1 minute. The java process also starts to collect data performance, with a 5 minute poll interval. All of the poll cycles are configurable.

The following table lists different PATROL for PostgreSQL collectors and explains how they work:

Collectors	Description
ConfigurationManager (15 sec)	Runs every 15 sec (default) and is responsible for checking if there was any change in the pconfig area.
ChannelListener (2 sec)	Responsible for opening a channel and starts a Java process and to listen to the java channel for incoming data.
AutoDiscovery (24 hour)	Run force discovery