troubleshoot-microsoft-sql-server

name	troubleshoot-microsoft-sql-server
description	Use when diagnosing issues with Microsoft SQL Server: log full / disk full, memory pressure spiral, worker thread exhaustion, tempdb contention, or blocking cascade. Queries Netdata via MCP for Microsoft SQL Server health signals, applies the diagnostic tree from the Netdata operator playbook, and recommends remediation.
version	0.1.0
author	Netdata
license	Apache-2.0
tags	["netdata","troubleshoot","mcp","microsoft-sql-server"]

Troubleshoot Microsoft SQL Server

When to use this skill

• Log full / disk full: Transaction log grows unchecked (missing log backups, long-running transaction, replication latency), fills the volume, halts all writes. The most common "unexpected" outage.
• Memory pressure spiral: Buffer pool shrinks (OS memory pressure, memory grants, max server memory misconfigured), PLE drops, physical I/O increases, I/O subsystem saturates, query latency rises, more concurrent sessions pile up, more memory pressure. Self-reinforcing.
• Worker thread exhaustion: Blocking chain or external wait causes workers to pile up. New connections get THREADPOOL waits. From the application's perspective, SQL Server stops responding.
• TempDB contention: Allocation page latch waits on PFS/GAM/SGAM pages in TempDB. Manifests as elevated PAGELATCH_UP / PAGELATCH_EX waits on pages in database ID 2. Common on high-concurrency OLTP.
• Blocking cascade: A single long-held lock (often an uncommitted transaction from an application bug, or a table lock from lock escalation) blocks dozens of other sessions. CPU looks idle, I/O looks idle, but nothing is processing.
• Parameter sniffing / bad plan: A query gets compiled with atypical parameter values, producing a catastrophically wrong plan. Can consume all memory grants, spill massively to TempDB, and saturate I/O.
• Any time the user reports a Microsoft SQL Server service behaving outside its expected envelope (elevated errors, latency, saturation, resource exhaustion, or unexpected restarts).
• An on-call engineer is paging on a Netdata alert tied to a Microsoft SQL Server instance and wants a structured triage path.

Key facts

• This skill wraps the Netdata operator playbook for Microsoft SQL Server. It does not replace the playbook; it routes a coding agent through MCP queries against the same signals the playbook relies on.
• Microsoft SQL Server is a single-process, multi-threaded relational database engine. The process (sqlservr.exe on Windows, sqlservr on Linux) manages its own memory, scheduling, and I/O subsystems through an internal layer called SQLOS; a user-mode operating system that sits between the SQL engine and the host OS.
• Dominant failure archetypes the playbook calls out: Log full / disk full; Memory pressure spiral; Worker thread exhaustion; TempDB contention; Blocking cascade.
• Netdata observes the signals listed in the rule files via its native collectors, plus any OpenTelemetry-shipped metrics that your Microsoft SQL Server instrumentation adds. Both paths end at the same MCP query surface.
• Netdata's mssql collector emits 71 context(s) under mssql.*. The rule files enumerate which contexts surface which domain; the Verification section below names the load-bearing ones explicitly.

Step-by-step

1. Confirm the Microsoft SQL Server service is up. Query Netdata via MCP with list_nodes and filter by the host running the target. A missing node means the symptom is at the network or orchestrator layer, not inside the service.
2. Pull the last 15 minutes of signals for the target. Use query_metrics against the contexts listed in the domain rule files. Run find_anomalous_metrics in parallel over the same window; anomalies frame which rule file to read first.
3. Check for Log full / disk full. Transaction log grows unchecked (missing log backups, long-running transaction, replication latency), fills the volume, halts all writes. The most common "unexpected" outage. Inspect the rule file whose signals move first for this mode.
4. Check for Memory pressure spiral. Buffer pool shrinks (OS memory pressure, memory grants, max server memory misconfigured), PLE drops, physical I/O increases, I/O subsystem saturates, query latency rises, more concurrent sessions pile up, more memory pressure. Self-reinforcing. Inspect the rule file whose signals move first for this mode.
5. Check for Worker thread exhaustion. Blocking chain or external wait causes workers to pile up. New connections get THREADPOOL waits. From the application's perspective, SQL Server stops responding. Inspect the rule file whose signals move first for this mode.
6. Check for TempDB contention. Allocation page latch waits on PFS/GAM/SGAM pages in TempDB. Manifests as elevated PAGELATCH_UP / PAGELATCH_EX waits on pages in database ID 2. Common on high-concurrency OLTP. Inspect the rule file whose signals move first for this mode.
7. Check for Blocking cascade. A single long-held lock (often an uncommitted transaction from an application bug, or a table lock from lock escalation) blocks dozens of other sessions. CPU looks idle, I/O looks idle, but nothing is processing. Inspect the rule file whose signals move first for this mode.
8. Correlate with host-level signals (system.cpu.utilization, system.memory.usage, system.disk.io_time). Many service-level failures have a host-resource precursor.
9. Apply the remediation hinted at in the matching rule file or the operator playbook. Re-run the MCP queries from the Verification section to confirm the signals returned to expected ranges. A fix that does not move the signal back is not a fix.

Handy MCP call templates

# Discover metrics from Microsoft SQL Server
list_metrics with q="mssql"

# Pull a specific context over the last window
query_metrics with context="mssql.user_connections", relative_window=-15m

# Rank anomalies for the service or host
find_anomalous_metrics with node=<host> and context_pattern="mssql.*"

# Correlate a known problem context with others
find_correlated_metrics around the incident window

# Show current alert state
list_raised_alerts scoped to the node

Common mistakes

• Treating Microsoft SQL Server as a generic HTTP or process health check. Microsoft SQL Server has specific failure archetypes (see Key facts) that generic checks miss.
• Stopping at the first anomalous metric. Several archetypes produce correlated spikes; use find_correlated_metrics to widen the search before concluding a root cause.
• Quoting percentile latency without the sample count. Low traffic plus a single slow request moves p99 by seconds.
• Reading dashboards for a window shorter than the failure's fingerprint. Slow-brew failures (queue growth, bloat, memory fragmentation) need 30+ minutes of data to see the trend.
• Skipping the host-level correlation. A process-level fix for a noisy-neighbour problem does not hold.
• Assuming alert thresholds are tuned for your workload. Tune against observed Microsoft SQL Server traffic before escalating an alert configuration issue.

Verification

Run these MCP queries against the Netdata instance that sees the Microsoft SQL Server service. Every context listed below is a real Netdata chart name; the agent does not need to guess.

1. list_metrics filtered by q="mssql" (returns every mssql.* context Netdata sees)
2. query_metrics with contexts=[mssql.user_connections, mssql.session_connections, mssql.buffer_page_lookups, mssql.memory_connection, mssql.database_backup_restore_throughput, mssql.database_read_only] and relative_window=-30m
3. find_anomalous_metrics filtered by node=<host> and context_pattern="mssql.*"

Load-bearing contexts for this service:

• mssql.user_connections: User Connections (connections). Dimensions: user.
• mssql.session_connections: Session Connections (connections). Dimensions: user, internal.
• mssql.buffer_page_lookups: Buffer Page Lookups (lookups/s). Dimensions: lookups.
• mssql.memory_connection: Connection Memory (bytes). Dimensions: memory.
• mssql.database_backup_restore_throughput: Backup/Restore Throughput (bytes/s). Dimensions: throughput.
• mssql.database_read_only: Database Read-Only Status (status). Dimensions: read_only, read_write.

A clean result means every context is within its expected band and the find_anomalous_metrics list is empty or contains only already-acknowledged items. If the fix was real, re-running the same queries 10 minutes after applying it will show a clean result. If it does not, revert and look deeper.

When the fix does not hold

If signals drift back into the anomalous range within 30 minutes of a remediation, the cause was deeper than the applied change. Typical misdiagnoses for Microsoft SQL Server:

• Host-resource pressure masquerading as application bug.
• Dependent service (DB, cache, upstream) causing a secondary symptom in the instrumented service.
• Configuration change that was never reloaded (some subsystems only pick up config on full restart).

Escalate by widening the query window: 2-6 hours instead of 15 minutes. Slow-moving causes are invisible at triage window sizes.

References

• rules/overview.md
• Netdata operator playbook: the authoritative source material this skill summarizes.
• skills/netdata-mcp-integration/ for the transport setup.
• skills/netdata-otel-setup/ if additional application signals are needed beyond what Netdata collects natively.