evidence-focused Cloud SQL Postgres evidence collection
Cloud SQL audit
Cloud SQL Postgres audit
Database Optimization Tool reviews Google Cloud SQL PostgreSQL evidence while keeping provider-managed limits separate from SQL-level findings.
Local collector runs from your environment
Human review before production decisions
No database changes applied by Database Optimization Tool
How Cloud SQL teams can start with a one-time audit
Cloud SQL Postgres can be audited from safe SQL evidence for many performance questions, but the report must account for managed settings and views the audit role cannot inspect.
Environment limits
Cloud SQL restricts host access and may limit extension control, privileged roles, and some settings. The audit avoids treating those hidden areas as verified facts.
What the report will not claim without more visibility
The collector can inspect relation size, index usage, table churn, vacuum signals, lock and activity snapshots, and statement statistics when the role can read them.
What this page cannot promise
Database Optimization Tool cannot change Cloud SQL flags, inspect Google Cloud-only metrics, or guarantee a finding when the database role lacks required view access.
FAQ
Frequently asked questions
These answers describe the product focus: careful database evidence, clear findings, and team-approved next steps.
Does a Cloud SQL audit require superuser?
No. A evidence-focused database user is the preferred model, and the report identifies any statistics or catalog access missing without superuser.
What if pg_stat_statements is unavailable on Cloud SQL?
The audit still reviews table, index, schema, and maintenance evidence. It will not pretend to rank query pressure fully without pg_stat_statements.
Will Database Optimization Tool change my Cloud SQL production database?
No. The collector gathers evidence-focused evidence and Database Optimization Tool does not apply database changes, change flags, or create indexes.
When should I read /docs/collector or /sample-report?
Read /docs/collector before exporting Cloud SQL evidence, and use /sample-report to see how limited visibility and recommendations are framed.