altinity-expert-clickhouse-part-log

Name: altinity-expert-clickhouse-part-log
Availability: InStock
Author: Altinity

Diagnose ClickHouse part_log activity across clusters with per-table and per-host granularity.

16|2|Updated Jan 8, 2026

Search Tags:#monitoring #troubleshooting #diagnostics #clickhouse #database-ops #part_log

Ready to run

npx skills add https://github.com/Altinity/Skills --skill altinity-expert-clickhouse-part-log

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: altinity-expert-clickhouse-part-log
Source: https://github.com/Altinity/Skills/tree/main/altinity-expert-clickhouse/skills/altinity-expert-clickhouse-part-log
Command: npx skills add https://github.com/Altinity/Skills --skill altinity-expert-clickhouse-part-log

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Part Log Based Diagnostics helps operators identify issues related to part creation and movement in ClickHouse, such as high NewPart rates, merge backlogs, and mutation storms, enabling faster diagnosis and remediation.

Core Features & Use Cases

Part_log wide analysis: cluster-wide checks of system.part_log to detect spikes and trends.
Per-table per-host visibility: surface offenders by database and table with event metrics.
Cross-module guidance: suggest loading related skills (ingestion, merges, replication, storage) when patterns appear.
Prebuilt checks: uses checks.sql for MCP-mode diagnostics.

Quick Start

Run the checks.sql queries against your cluster to surface top offenders and guide remediation actions.

Frequently Asked Questions about altinity-expert-clickhouse-part-log

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I diagnose high part churn and merge backlogs in ClickHouse?▼

Diagnosing part churn in ClickHouse involves analyzing system.part_log activity cluster-wide. This Skill uses time-bounded queries via clusterAllReplicas to detect high NewPart rates, slow merges, and mutation storms, surfacing top offending tables and hosts.

What causes mutation storms in ClickHouse and how do I detect them?▼

Mutation storms in ClickHouse are detected by analyzing system.part_log events for spikes in mutation activity. This Skill applies cluster-wide diagnostics with version-aware schemas to identify mutation storms and guide remediation across affected databases and tables.

Can I monitor ClickHouse part_log activity across all replicas in a cluster?▼

Yes, you can monitor ClickHouse part_log activity across all replicas using clusterAllReplicas. This Skill performs cluster-wide analysis with per-table and per-host granularity to surface top offenders and detect trends in part creation and movement.

What's the best way to find which ClickHouse tables are generating the most parts?▼

The best way to find tables generating the most parts in ClickHouse is running part_log diagnostics with per-table and per-host visibility. This Skill surfaces top offenders by database and table using event metrics from time-bounded clusterAllReplicas queries.

How do I troubleshoot slow merges in a ClickHouse cluster?▼

To troubleshoot slow merges in a ClickHouse cluster, query system.part_log to detect merge backlogs and identify affected tables and hosts. This Skill applies deterministic time-bounded analysis with version-aware schemas to surface slow merge patterns and guide remediation.