The optimal value for Max Worker Threads is the default, 0, which tells SQL Server to calculate the limit from the number of logical processors. This post shows how to read the real limit, how the formula works, and which wait and queue counters prove a thread shortage before you change anything.
Collecting Server Hardware Facts Before a Tuning Session
Capture server hardware facts from SQL Server views, keep an inventory, and distinguish visible resources from workload sizing evidence.
High CPU Consumption in SQL Server: Where to Look First
High CPU consumption in SQL Server comes from a handful of causes, and a short checklist finds them in order. This post reads the scheduler queue, ranks cached queries by CPU, and checks the Common Criteria Compliance setting that caused a known CPU problem after upgrades. It ends with the update that fixed that case and the questions to ask before you add cores.
CPU-Bound Queries: Comparing worker_time to elapsed_time
Is a slow query working or waiting? Compare its two clocks and you will know what to fix before anyone buys more CPU.
Grouping by query_hash to Find Cheap Queries That Cost the Most
Group cached statements by query_hash to find high total CPU and reads, compare plan hashes, and account for missing cache history.
SQL SERVER 2016 – DBCC Command Enhancement with MAXDOP
SQL Server 2016 lets you set MAXDOP on DBCC checks like CHECKDB. Here is how this DBCC Command enhancement works and how to use it.
SQL SERVER – Inspecting Historical CPU Ring Buffer Samples
CPU Usage History helps me establish when a reported problem began. It gives a timeline rather than identifying the responsible query.