Troubleshooting overview

This page summarizes common CloudBase PostgreSQL troubleshooting paths. First identify whether the issue is related to connection, permissions, SQL execution, HTTP API, backup and recovery, or performance.

Quick diagnosis

Symptom	Start here
Unsure where to start, or seeing common HTTP 401/403/404/409/500 errors	Common errors
Cannot connect, timeout, wrong password	Connection failures
Empty query result or permission denied	Permission and RLS issues
New HTTP API field is missing or RPC function is not found	Schema cache issues
Backup failed or restored data is unexpected	Backup and restore issues
Slow queries, queued pool, lock waits	Query optimization and Logs and monitoring

Collect information

Before self-diagnosis or reporting an issue, collect:

• Environment ID, region, database instance, and table name.
• Access method: SDK, HTTP API, DMC, Cloud Functions, CloudBase Run, or direct connection.
• Complete error message and SQLSTATE code.
• Request time, request ID, user identity, and key parameters.
• Recent changes to schema, RLS policies, role grants, or connection configuration.

Useful SQL checks

Check the current connection:

select current_database(), current_user, now();

Check whether tables exist:

select table_schema, table_name
from information_schema.tables
where table_schema = 'public'
order by table_name;

Check whether RLS is enabled:

select schemaname, tablename, rowsecurity
from pg_tables
where schemaname = 'public';

Handling principles

• Reproduce in a test environment before changing production.
• Preserve evidence before rollback, restart, or cleanup.
• Keep SQL records for permission and data fixes.
• For issues with unclear impact, pause high-risk writes first.