Søren Eggert Lundsteen Olsen
2284e7f082
A DBA should be able to review every statement SQLInsight sends to a monitored server before granting it access. "SQL Scripts/<rule-id>.sql" holds, per rule (82 built-in, 16 community), a Danish header (what and why, how C# judges the data, thresholds with config keys, severity, exact GRANTs, versions, collector file and member) and the collector's SQL verbatim, every version variant as its own batch. "_collectors/<Collector>.sql" holds each collector's queries once. The only statement that is not strictly read-only is documented in the scripts: the deadlock collector's session temp table (#system_health). Inline SQL in the collectors moved to internal constants and Build*Sql methods (no text changed), so SqlScriptTests can check that the scripts still contain the SQL the code sends, that every rule has a script, that every collector query is registered, and that nothing writes. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
77 lines
4.7 KiB
Transact-SQL
77 lines
4.7 KiB
Transact-SQL
-- ===================================================================================================================
|
|
-- SQLI-CPU-001: Vedvarende højt CPU-forbrug
|
|
-- ===================================================================================================================
|
|
-- Regel-id: SQLI-CPU-001
|
|
-- Titel: Vedvarende højt CPU-forbrug
|
|
-- Alvor: SQL Server: Medium fra 75 %, High fra 90 %. Andre processer: Medium.
|
|
-- Versioner: SQL Server 2005 og nyere. "Andre processer" kan ikke måles på Linux (host_platform fra SQL Server
|
|
-- 2017) og Azure SQL Managed Instance.
|
|
--
|
|
-- Hvad tjekkes, og hvorfor:
|
|
-- Om CPU'en har været vedvarende høj, opdelt i SQL Servers eget forbrug og andre processers.
|
|
--
|
|
-- Vurdering (sker i C# ud fra de indsamlede data, ikke i SQL'en, medmindre andet er nævnt):
|
|
-- Scheduler monitor skriver én post i minuttet (ProcessUtilization = SQL Server, SystemIdle = ledig). Reglen bruger
|
|
-- de seneste op til 15 poster (kræver mindst 5). Fund 1: gennemsnitlig ProcessUtilization >= 75 %. Fund 2:
|
|
-- gennemsnitlig andel til andre processer (100 - SystemIdle - ProcessUtilization) >= 25 %.
|
|
--
|
|
-- Grænseværdier (konfigurationsnøgle i agentens appsettings.json og standardværdi):
|
|
-- 15 minutter (CpuRule.Minutes); 75 % / 90 % for SQL Server; 25 % for andre (fast i koden).
|
|
--
|
|
-- Rettigheder til agentens login (se db/permissions/grant-least-privilege.sql; [agent] = agentens login):
|
|
-- GRANT VIEW SERVER STATE TO [agent]; -- SQL Server 2005-2019
|
|
-- GRANT VIEW SERVER PERFORMANCE STATE TO [agent]; -- SQL Server 2022+ (erstatter VIEW SERVER STATE)
|
|
-- GRANT VIEW SERVER SECURITY STATE TO [agent]; -- SQL Server 2022+
|
|
--
|
|
-- Data kommer fra (collector, fil : medlemmer, interval):
|
|
-- CpuCollector
|
|
-- Fil: src/SQLInsight.Collectors/Performance/CpuCollector.cs
|
|
-- Medlemmer: CpuCollector.Sql, CpuCollector.HostPlatformSql
|
|
-- Interval: 5 min som standard (SQLInsight:CollectorIntervals:cpu)
|
|
-- Delt script: SQL Scripts/_collectors/CpuCollector.sql
|
|
-- Regelkode: src/SQLInsight.Rules/Performance/ResourceRules.cs : CpuRule.Evaluate
|
|
--
|
|
-- Read-only: Ja: kun SELECT samt lokale variabler, SET-sessionsindstillinger. Intet ændrer data, indstillinger
|
|
-- eller objekter.
|
|
-- Kørsel: Kør i SSMS som agentens login (eller et login med samme rettigheder). Hver variant er sin egen batch
|
|
-- (GO); en variant for en anden SQL Server-version end din kan fejle - det er forventet.
|
|
-- Per-database-forespørgsler er vist for databasen msdb, som findes overalt: ret navnet til den
|
|
-- database, du vil se.
|
|
-- ===================================================================================================================
|
|
|
|
-- -------------------------------------------------------------------------------------------------------------------
|
|
-- Forbindelsen: Sessionsindstillinger, agenten sætter på hver forbindelse
|
|
-- -------------------------------------------------------------------------------------------------------------------
|
|
|
|
-- Variant: Alle versioner: sættes på hver forbindelse, agenten åbner
|
|
-- Kilde: SqlConnectionFactory.SessionSettings
|
|
-- Kode: src/SQLInsight.Connectivity/SqlConnectionFactory.cs : SqlConnectionFactory.SessionSettings
|
|
SET LOCK_TIMEOUT 5000; SET DEADLOCK_PRIORITY LOW; SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
|
|
GO
|
|
|
|
-- -------------------------------------------------------------------------------------------------------------------
|
|
-- CpuCollector: SQL Servers og systemets CPU-forbrug pr. minut (scheduler monitor ring buffer, op til 60 minutter).
|
|
-- -------------------------------------------------------------------------------------------------------------------
|
|
|
|
-- Variant: SQL Server 2005 og nyere
|
|
-- Kilde: CpuCollector.Sql
|
|
-- Kode: src/SQLInsight.Collectors/Performance/CpuCollector.cs : CpuCollector.Sql
|
|
DECLARE @now_ticks bigint;
|
|
SELECT @now_ticks = ms_ticks FROM sys.dm_os_sys_info;
|
|
SELECT TOP (60)
|
|
DATEADD(second, CAST((rb.[timestamp] - @now_ticks) / 1000 AS int), GETUTCDATE()),
|
|
rb.record.value('(./Record/SchedulerMonitorEvent/SystemHealth/ProcessUtilization)[1]', 'int'),
|
|
rb.record.value('(./Record/SchedulerMonitorEvent/SystemHealth/SystemIdle)[1]', 'int')
|
|
FROM (
|
|
SELECT [timestamp], CONVERT(xml, record) AS record
|
|
FROM sys.dm_os_ring_buffers
|
|
WHERE ring_buffer_type = N'RING_BUFFER_SCHEDULER_MONITOR' AND record LIKE N'%<SystemHealth>%'
|
|
) rb
|
|
ORDER BY rb.[timestamp] DESC
|
|
GO
|
|
|
|
-- Variant: SQL Server 2017 og nyere (Windows eller Linux)
|
|
-- Kilde: CpuCollector.HostPlatformSql
|
|
-- Kode: src/SQLInsight.Collectors/Performance/CpuCollector.cs : CpuCollector.HostPlatformSql
|
|
SELECT host_platform FROM sys.dm_os_host_info
|
|
GO
|