Azure PostgreSQL Lesson Learned #8: Post-Upgrade Performance Surprises (The One-Step Fix)
Co‑authored with angesalsaa Symptoms Upgrade from PostgreSQL 12 → higher version succeeds. After migration, workloads show: Queries running slower than before. Unexpected CPU spikes during normal operations. No obvious errors in logs or connectivity issues. Root Cause Missing or stale statistics can lead to bad query plans, which in turn might degrade performance and consume excessive memory. After a major version upgrade, the query planner relies on outdated or default estimates because the optimizer’s learned patterns are not refreshed. This often results in: Sequential scans instead of index scans. Inefficient join strategies. Increased CPU and memory usage. Contributing Factors Large tables with skewed data distributions. Complex queries with multiple joins. Workloads dependent on accurate cost estimates. Specific Conditions We Observed Any source server version can be impacted once you upgrade to higher version. No ANALYZE or VACUUM run post-upgrade. Operational Checks Before troubleshooting, confirm: Query plans differ significantly from pre-upgrade. pg_stats indicates outdated or missing statistics. Mitigation Goal: Refresh statistics so the planner can optimize queries. Run ANALYZE on all tables: ANALYZE; Important Notes: These commands are safe and online. For very large datasets, consider running during low-traffic windows. We recommend running the ANALYZE command in each database to refresh the pg_statistic table. Post-Resolution Queries return to expected performance. CPU utilization stabilizes. Execution plans align with indexes and cost-based optimization. Prevention & Best Practices Always schedule ANALYZE immediately after major upgrades. Automate stats refresh in your upgrade runbook. Validate plans for critical queries before going live. Why This Matters Skipping this step can lead to: Hours of degraded performance. Emergency escalations and customer dissatisfaction. Misdiagnosis as engine regression when it’s just missing stats. Key Takeaways Issue: Post-upgrade query slowness and CPU spikes due to stale/missing statistics. Fix: Run ANALYZE immediately after upgrade. Pro Tip: Automate this in CI/CD or maintenance scripts for zero surprises. References Major Version Upgrades - Azure Database for PostgreSQL | Microsoft LearnLesson Learned #489:Investigating CPU Spikes with Query Store Overall Resource Consumption Report
Yesterday, I worked on a customer who experienced a CPU spike that went 90% during the timeframe of 19:00 to 21:00. I would like to share how we identified the queries that potentially caused this CPU spike using the Query Store's Overall Resource Consumption report.
