How would you use a Redshift probe to monitor and troubleshoot performance issues?

A Redshift probe is a lightweight query that inspects system tables to surface performance bottlenecks. To use it, first run the probe against the SVV_QUERY_METRICS and SVV_WLM_QUERY tables to get metrics like query duration, CPU usage, and queue time. Next, examine the SVV_QUERY_PLAN table to see the execution plan and identify scans, joins, or sorts that are expensive. The probe can also query STV_RECENTS to see the most recent queries and their resource consumption. If you notice high queue times, check the WLM configuration (SVV_WLM_CONFIG) to adjust queue sizes or concurrency scaling. For slow scans, look at SVV_TABLE_INFO to see if tables are properly distributed and if sort keys are used. The probe can be scheduled as a cron job or integrated into CloudWatch alarms. By correlating probe results with query logs, you can pinpoint whether the issue is due to data skew, missing statistics, or insufficient memory. Once identified, you can take corrective actions such as vacuuming, re‑distributing, or adding sort keys. This systematic probing approach helps maintain optimal Redshift performance.