Health assessment reports provide a point-in-time snapshot of the performance and health of a PG cluster node over a given period. You can use them to identify anomalies, track how your system evolves over time, and share a consistent artifact with others.
Reports are generated automatically: one report per node immediately after a cluster is connected to Hybrid Manager (HM), then once every 24 hours thereafter. All nodes are tracked, including witness nodes — reports aren't scoped to a cluster or data group. Reports are stored persistently, so they remain available even after the underlying data used to generate them has been purged.
Report contents
Each report covers a fixed time window and includes aggregate statistics for that period, such as session counts, transaction stats, wait events, top SQL statements, WAL and buffer metrics.
Viewing and downloading reports
To access reports:
- In the HM console, select Reports.
The Reports page lists all reports across your estate, sorted by date. You can filter by project, cluster, date range, and report interval.
- To download a report, select the report row and select Download Report. Reports are available as PDF files.
You can also view reports scoped to a single cluster from the cluster detail page.
Comparing two reports
You can ask Chat Agent to compare two reports for the same cluster. This highlights the changes between two snapshots — useful for tracking how performance has changed over time.
To compare reports using Chat Agent:
- Open Chat Agent in HM.
- Ask Chat Agent to compare reports for a specific node and cluster — for example: "Compare the last two reports for the primary node of cluster X."
Chat Agent returns a summary of the differences between the two reports.
Note
To use the chatbot comparison feature, Chat Agent must be enabled in HM and the model must be configured as at least a Large (L) model. See EDB Hybrid Manager Chat Agent for the instructions.