PostgreSQL Memory helps review memory-related metrics and trends for a monitored PostgreSQL server. Use it when workload behavior suggests memory pressure or configuration issues.
Workflow
- Open Memory.
- Review current and recent memory values.
- Compare with Activity and Queries.
- Check Settings for relevant PostgreSQL configuration values.
- Use Health Checks for broader configuration findings.
How To Review PostgreSQL Memory
PostgreSQL memory evidence helps explain whether a workload is constrained by cache, sort/hash operations, connection count, or host pressure. Use this page when many queries become slower at the same time, when I/O increases unexpectedly, or when health checks identify configuration concerns.
Review memory alongside workload. A high connection count can multiply per-session memory use, while a small number of large analytical queries can create very different pressure. If reads are high at the same time, compare with PostgreSQL I/O. If the problem is tied to specific statements, use PostgreSQL Queries before changing server-wide memory settings.
Avoid tuning PostgreSQL memory from a single snapshot. Capture the affected time window, the active queries, wait patterns, and host metrics where available. Then make controlled changes and compare before-and-after behavior.
PostgreSQL Memory FAQ
Should I increase memory settings first?
Not automatically. Query design, indexes, connection count, and I/O pressure can all create memory symptoms.
Why does connection count matter?
Some PostgreSQL memory is allocated per session or operation. Many active sessions can increase total demand.
Where should I check configuration values?
Use PostgreSQL Settings and PostgreSQL Health Checks for configuration context.
Next Steps From Memory Evidence
After reviewing memory evidence, decide whether the pressure is query-specific, connection-related, or host-wide. Query-specific pressure should be compared with Queries and Indexes. Connection-related pressure should be compared with Activity and application pool settings. Host-wide pressure needs host OS metrics where available. Make configuration changes only after collecting a baseline.