Monitor and tune Azure Database for PostgreSQL - Single Server

APPLIES TO: Azure Database for PostgreSQL - Single Server

Monitoring data about your servers helps you troubleshoot and optimize for your workload. Azure Database for PostgreSQL provides various monitoring options to provide insight into the behavior of your server.


Azure Database for PostgreSQL provides various metrics that give insight into the behavior of the resources supporting the PostgreSQL server. Each metric is emitted at a one-minute frequency, and has up to 93 days of history. You can configure alerts on the metrics. For step by step guidance, see How to set up alerts. Other tasks include setting up automated actions, performing advanced analytics, and archiving history. For more information, see the Azure Metrics Overview.

List of metrics

These metrics are available for Azure Database for PostgreSQL:

Metric Metric Display Name Unit Description
cpu_percent CPU percent Percent The percentage of CPU in use.
memory_percent Memory percent Percent The percentage of memory in use.
io_consumption_percent IO percent Percent The percentage of IO in use. (Not applicable for Basic tier servers.)
storage_percent Storage percentage Percent The percentage of storage used out of the server's maximum.
storage_used Storage used Bytes The amount of storage in use. The storage used by the service may include the database files, transaction logs, and the server logs.
storage_limit Storage limit Bytes The maximum storage for this server.
serverlog_storage_percent Server Log storage percent Percent The percentage of server log storage used out of the server's maximum server log storage.
serverlog_storage_usage Server Log storage used Bytes The amount of server log storage in use.
serverlog_storage_limit Server Log storage limit Bytes The maximum server log storage for this server.
active_connections Active Connections Count The number of active connections to the server.
connections_failed Failed Connections Count The number of established connections that failed.
network_bytes_egress Network Out Bytes Network Out across active connections.
network_bytes_ingress Network In Bytes Network In across active connections.
backup_storage_used Backup Storage Used Bytes The amount of backup storage used. This metric represents the sum of storage consumed by all the full database backups, differential backups, and log backups retained based on the backup retention period set for the server. The frequency of the backups is service managed and explained in the concepts article. For geo-redundant storage, backup storage usage is twice that of the locally redundant storage.
pg_replica_log_delay_in_bytes Max Lag Across Replicas Bytes The lag in bytes between the primary and the most-lagging replica. This metric is available on the primary server only.
pg_replica_log_delay_in_seconds Replica Lag Seconds The time since the last replayed transaction. This metric is available for replica servers only.

Server logs

You can enable logging on your server. These resource logs can be sent to Azure Monitor logs, Event Hubs, and a Storage Account. To learn more about logging, visit the server logs page.

Query Store

Query Store keeps track of query performance over time including query runtime statistics and wait events. The feature persists query runtime performance information in a system database named azure_sys under the query_store schema. You can control the collection and storage of data via various configuration knobs.

Query Performance Insight

Query Performance Insight works in conjunction with Query Store to provide visualizations accessible from the Azure portal. These charts enable you to identify key queries that impact performance. Query Performance Insight is accessible from the Support + troubleshooting section of your Azure Database for PostgreSQL server's portal page.

Performance Recommendations

The Performance Recommendations feature identifies opportunities to improve workload performance. Performance Recommendations provides you with recommendations for creating new indexes that have the potential to improve the performance of your workloads. To produce index recommendations, the feature takes into consideration various database characteristics, including its schema and the workload as reported by Query Store. After implementing any performance recommendation, customers should test performance to evaluate the impact of those changes.

Planned maintenance notification

Planned maintenance notifications allow you to receive alerts for upcoming planned maintenance to your Azure Database for PostgreSQL - Single Server. These notifications are integrated with Service Health's planned maintenance and allow you to view all scheduled maintenance for your subscriptions in one place. It also helps to scale the notification to the right audiences for different resource groups, as you may have different contacts responsible for different resources. You will receive the notification about the upcoming maintenance 72 hours before the event.

Learn more about how to set up notifications in the planned maintenance notifications document.

Next steps