Practical articles about ClickHouse, SQL Server, analytics architecture, performance, migration, and database operations.
A practical guide to finding parameter-sensitive query plans in SQL Server, proving the cause, and choosing the least risky fix.
Read articleA job can report success while skipping databases, processing no rows, or finishing too late for downstream work. Final status alone does not prove the intended result.
Read articleTemporary tables and table variables both hold intermediate rows, but statistics, compilation, indexing, and transaction behavior can produce very different execution plans.
Read articleRow and page compression can reduce storage and I/O, but add CPU work and do not benefit every table equally. The decision depends on data shape and workload.
Read articlePartitioning is often presented as an automatic performance feature. Its strongest benefits usually involve lifecycle, sliding-window loads, archival, and maintenance boundaries.
Read articleDynamic SQL is useful for optional filters and variable query shapes, but string concatenation introduces injection risk, quoting defects, poor plan reuse, and difficult troubleshooting.
Read articleStorage, memory, firmware, and software failures can damage pages without immediately stopping the application. Backups can preserve corruption that was never detected.
Read articleReaders and writers can block each other under lock-based read committed isolation. RCSI uses row versions to reduce that blocking, but changes resource use and concurrency behavior.
Read articleAn index cannot help when a predicate hides its column inside a function or calculation. SQL Server may scan every candidate row instead of navigating to a range.
Read articleWhen compared values use incompatible data types, SQL Server may convert an indexed column at runtime. The result can be a scan, poor estimates, and unnecessary CPU.
Read articleUnexpected transaction log growth can fill a volume, slow transactions, delay replicas, and extend recovery. Shrinking the log treats a symptom and often repeats the cycle.
Read articleLeaving max server memory at its default can starve Windows and co-located services. Setting it too low causes avoidable reads, compilations, and memory-grant pressure.
Read articleDefault parallelism settings are rarely an intentional workload design. Excessive parallelism can amplify CPU and worker pressure, while restrictive settings can slow analytical work.
Read articleA long list of blocked sessions can make every query look guilty. Usually one transaction near the front of the chain controls the incident's duration and impact.
Read articleA successful backup job proves that SQL Server wrote a backup stream. It does not prove that the organization can restore the right databases, keys, log chain, and dependencies on time.
Read articleAn Availability Group can show healthy replicas while recovery objectives remain at risk. Queue growth, listener behavior, backup placement, and failover readiness require explicit monitoring.
Read articleThe optimizer chooses joins, access methods, and memory grants from estimated row counts. When estimates diverge from reality, a reasonable plan can become expensive.
Read articleTempDB supports sorts, hashes, row versioning, temporary objects, spills, and internal worktables. Poor layout or slow storage can turn it into a shared bottleneck.
Read articleA deadlock is a cycle in which sessions hold resources needed by one another. SQL Server terminates one participant, but the victim is not necessarily the root cause.
Read articleMany performance incidents disappear before an engineer captures the plan. Query Store preserves query, plan, runtime, and wait history for later analysis.
Read articleRebuilding every index on a fixed schedule consumes CPU, I/O, log space, and maintenance time even when users receive no measurable benefit.
Read article