Azure SQL Database insights

High sqlserver_page_split_rate indicates fragmented indexes or poor fill-factor settings, causing additional I/O and transaction log writes during insert/update operations.

Low sqlserver_page_buffer_cache_hit_ratio combined with high sqlserver_page_lazy_write_rate and declining sqlserver_page_life_expectancy indicates memory pressure, forcing excessive physical I/O and degrading query performance.

During dynamic reconfiguration events (software upgrades, load balancing, failovers), transient connection failures (errors 40197, 40613) can exhaust connection pools if retry logic is absent or poorly tuned, leading to cascading failures.

Query plans with missing indexes, parameter sniffing issues, or non-sargable predicates consume excessive CPU, leading to DTU saturation or query timeouts. This often manifests as sustained high sqlserver_batch_request_rate with elevated user CPU time.

Elevated sqlserver_lock_wait_rate and sqlserver_lock_wait_time_avg signal blocking chains or deadlocks, preventing transactions from completing and causing query timeouts or application-level failures.

When the transaction log cannot be truncated due to long-running or blocked transactions, the log file grows until it hits maximum size limits, causing errors 9002 or 40552 and preventing new write transactions.

Sudden drops in sqlserver_transaction_rate or sqlserver_transaction_write_rate, when not correlated with reduced application load, may indicate resource exhaustion, blocking, or configuration changes affecting commit throughput.

Database Performance Bottlenecks Invisible Without Query Insights

SQL database performance issues remain undiagnosed without Query Performance Insights enabled, as high-level CPU and memory metrics don't reveal slow queries, missing indexes, or inefficient execution plans.