Metrics and performance
Watch a managed database's CPU, memory, storage, connections and queries, see live sessions and the most expensive statements, and spot unused indexes.
Usage at a glance
A database's Overview shows its CPU, memory and storage use against its plan, updated every minute. Storage beyond the plan is billed per GB-month; the Overview says how much and suggests a larger plan.
Metrics
The Metrics tab charts connections, queries and traffic, sampled every 5 minutes:
- Connections: open client connections, against the number allowed;
- Queries / sec and Transactions / sec;
- Cache hit rate: reads served from memory;
- Traffic in and out;
- Resources: CPU, memory and storage against the plan, with CPU and memory live for the last 15 minutes.
Graphs appear about 10 minutes after the database starts running.
The Metrics tab also shows:
- Sessions: clients connected right now, with their state and query. Cancel query stops a statement (the session stays connected); End session disconnects the client and rolls back its open transaction.
- Configuration: the engine's settings (read-only).
Performance Insights
Performance Insights shows, live from the database server, what is running and what costs the most. It covers all databases on the server by default; pick one to see its tables, indexes and statements.
- What's running now: active queries; those running longer than 30 seconds are highlighted. To cancel one, use Sessions on the Metrics tab.
- Busiest and largest databases on the server.
- Top queries: the statements that cost the most, with values replaced by
?, sorted by Total time, Per run or Most runs. Statements read without an index are flagged. - Index health: indexes that are never used, or already covered by another one. They slow down writes and use disk.
Related
Data Explorer and SQL console
Browse and edit tables in the browser, run SQL safely in read-only or write mode, import and export data, track schema changes and run one script on many databases.
Object Storage
Store files in S3-compatible buckets and reach them with the S3 tools you already use, from the AWS CLI and SDKs to rclone and Terraform.