How do I troubleshoot performance on our self-hosted Cruise database?

This article provides operational boundaries, maintenance requirements, and diagnostic steps for self-hosted database instances, including ship and select corporate environments.

Roles & Support Boundaries

Managing a self-hosted database requires clear division of operational responsibilities between your organization and Crunchtime Support.

Customer Responsibility

Self-hosted customers are strictly responsible for monitoring, administering, maintaining, and troubleshooting their own database performance. Customer Database Administrators (DBA) must regularly gather database statistics, perform index maintenance, and tune Oracle configurations to standard specifications.

Crunchtime Support Scope

While Crunchtime can make recommendations and provide diagnostic suggestions as a courtesy, we do not provide ongoing administration, continuous monitoring, or database tuning for self-hosted environments. Deeper configuration issues should be addressed with Oracle Support. Support engagement depth may vary for clients enrolled in Platinum Support.

Hardware & Infrastructure Requirements

Ensure the database server meets the Cruise Minimum System Specifications. Underpowered CPU, insufficient RAM, slow disk I/O, or virtual machine resource competition are frequent root causes of transaction stalls.

How to Perform Customer Pre-Ticket Diagnostics

Before submitting a ticket to Crunchtime Support, customer DBAs should complete these preliminary diagnostic steps to identify and resolve common performance bottlenecks.

Verify Server Health. Check server CPU, memory, and disk I/O utilization to confirm they are within normal operating thresholds. Verify that no virtual machine resource contention is occurring.
Check Utility Error Logs. Query T_UTL_ERR_LOG under the IM_OWNER schema to identify system-level execution errors recorded during the past 7 days with the following SQL query:
```
select * from t_utl_err_log
where err_dttm > sysdate - 7
order by err_dttm desc;
```
Regather Stale Statistics. Outdated statistics lead to suboptimal execution plans. Have your DBA re-analyze schema statistics, ensuring they are up-to-date.
Identify Active Application Sessions & Slow Queries. Monitor active connections and slow queries associated with the application user account by querying v$session.

Get Additional Help from Crunchtime Support

If database performance issues persist after completing basic maintenance and preliminary checks, reach out to Crunchtime Support for further guidance.

What Information Does Crunchtime Support Need to Help Troubleshoot?

When submitting a support ticket, provide the following details to assist Crunchtime Support in analyzing the issue efficiently:

Scope of Impact: Indicate whether the issue affects a single location/ship, multiple locations, or is fleetwide.
Reproduction Steps: Provide the specific user action, report, or transaction stalling in the application, along with exact timestamps.
SQL Diagnostics: Include identified SQL_ID(s) and execution statistics gathered during your initial investigation.
Up-to-Date Statistics Confirmation: Confirm that schema and table statistics have been re-gathered and are up to date.
Error Logs: Provide any relevant error logs recorded during the issue, including query results from T_UTL_ERR_LOG.