ClickHouse analytics DB

Install

emisar pack install fetches this pack, re-validates it, and verifies its content hash against the --hash below — the exact bytes this page was rendered against, so a tampered copy is rejected — before copying it into the runner's packs dir. The command reloads a running daemon itself; no manual restart.

sudo emisar pack install clickhouse --hash sha256:102855cbb046d9b72e7297e749c39d0f787dbc6fd237511c71a59a46eef366a9

• ch.asynchronous_metrics exec low

  SELECT * FROM system.asynchronous_metrics

  Show periodically-computed metrics — replica lag (ReplicasMaxAbsoluteDelay), max parts per partition, filesystem and memory. The async counterpart to system.metrics.

  View source on GitHub
• ch.backups exec low

  SELECT * FROM system.backups

  List BACKUP/RESTORE operations since the last server start with status and error. Non-persistent across restarts.

  View source on GitHub
• ch.detached_parts exec low

  SELECT * FROM system.detached_parts

  List parts ClickHouse detached and will not auto-attach (corruption, manual detach, broken), each with a reason. Empty is healthy.

  View source on GitHub
• ch.dictionaries exec low

  SELECT * FROM system.dictionaries

  List loaded dictionaries with status, element count, source.

  View source on GitHub
• ch.disks exec low

  SELECT * FROM system.disks

  List configured disks with free/used bytes + paths.

  View source on GitHub
• ch.distribution_queue exec low

  SELECT * FROM system.distribution_queue

  Show distributed-table async send backlog — pending files, errors, blocked sends. Spot a Distributed table not flushing to shards.

  View source on GitHub
• ch.errors exec low

  SELECT * FROM system.errors

  List error codes seen since startup with counts and the last message. Spot error storms.

  View source on GitHub
• ch.keeper_status exec low

  List a Keeper/ZooKeeper path (system.zookeeper)

  List child nodes and metadata under a Keeper/ZooKeeper path. system.zookeeper requires a path filter, so the path arg is mandatory.

  View source on GitHub
• ch.kill_mutation exec critical

  Cancel a stuck mutation (KILL MUTATION)

  Cancel one ClickHouse mutation by its (database, table, mutation_id) — KILL MUTATION WHERE …. Use to stop a mutation wedged on a failure (see ch.stuck_mutations) so the table's merges and inserts can proceed. Destructive — the mutation is abandoned mid-flight — so risk:critical → default-denied. An operator allows it by explicit policy, and WHICH mutations are permitted is a portal policy decision.

  View source on GitHub
• ch.kill_query exec high

  KILL QUERY WHERE query_id = '<id>'

  Cancels one running query by ID.

  View source on GitHub
• ch.merge_queue exec low

  SELECT * FROM system.merges

  List in-progress + queued merges.

  View source on GitHub
• ch.mutation_queue exec low

  SELECT * FROM system.mutations

  List in-progress + pending mutations (ALTER TABLE).

  View source on GitHub
• ch.optimize_table exec high

  OPTIMIZE TABLE FINAL

  Forces merge of all parts into one. Heavy disk + CPU.

  View source on GitHub
• ch.parts_by_partition exec low

  Active part count per partition (system.parts)

  List active parts grouped by partition, top offenders first. Where you actually diagnose too-many-parts / merge backlog (parts_summary is table-level only).

  View source on GitHub
• ch.parts_summary exec low

  SELECT system.parts (by table)

  List active part counts + sizes per table. Use to spot too-many-parts.

  View source on GitHub
• ch.processes exec low

  SELECT * FROM system.processes

  List currently-running queries.

  View source on GitHub
• ch.query_errors exec low

  Failed queries from system.query_log (last hour)

  List queries that failed in the last hour (exception types 3 and 4) with code, message, and query. The error counterpart to slow_queries.

  View source on GitHub
• ch.reload_config exec high

  SYSTEM RELOAD CONFIG

  Re-reads config.xml + users.xml without restarting.

  View source on GitHub
• ch.replicas_status exec low

  SELECT * FROM system.replicas

  Show per-replica state — leader, queue size, log delays, errors.

  View source on GitHub
• ch.replication_queue exec low

  SELECT * FROM system.replication_queue

  Show per-task replication queue — retries, postpones, last exception. Spot stuck fetches/merges (distinct from per-replica state).

  View source on GitHub
• ch.slow_queries exec low

  system.query_log slowest queries

  List top 50 slowest queries from the last hour.

  View source on GitHub
• ch.stuck_mutations exec low

  Failing mutations (system.mutations)

  List mutations stuck on a failure (latest_fail_reason set), with the failed part and reason. Narrower than the full mutation queue — surfaces only what is actively failing.

  View source on GitHub
• ch.system_drop_replica exec critical

  SYSTEM DROP REPLICA <name>

  Remove a replica's metadata from ZooKeeper for one table. Use only for an already-dead replica that is never coming back. The replica cannot rejoin without recreating its tables. Wrong replica name destroys a healthy cluster member's metadata.

  View source on GitHub
• ch.system_events exec low

  SELECT * FROM system.events

  List process-lifetime event counters.

  View source on GitHub
• ch.system_flush_logs exec medium

  SYSTEM FLUSH LOGS

  Force write buffered system.*log tables to disk. Use before querying very recent activity from query_log, part_log, etc.

  View source on GitHub
• ch.system_metrics exec low

  SELECT * FROM system.metrics

  Show per-metric current values (gauges + counters).

  View source on GitHub
• ch.system_restart_replica exec high

  SYSTEM RESTART REPLICA <table>

  Reinitialize the local replication state for one table. Useful when the replica is stuck or its ZooKeeper state diverged. Heavy — may re-fetch many parts.

  View source on GitHub
• ch.system_sync_replica exec medium

  SYSTEM SYNC REPLICA <table>

  Wait for the local replica to catch up with all peers. Returns when in sync. Use to confirm replication caught up before doing reads from this replica.

  View source on GitHub
• ch.table_sizes exec low

  system.tables uncompressed sizes

  List top 50 tables by uncompressed size.

  View source on GitHub
• ch.tables_overview exec low

  Engine inventory (system.tables)

  List all non-system tables with engine, row/byte totals, and part counts. Engine inventory + fleet shape.

  View source on GitHub
• ch.uptime exec low

  SELECT version, uptime

  Show server version + uptime.

  View source on GitHub
• ch.zookeeper_connection exec low

  SELECT * FROM system.zookeeper_connection

  List active Keeper/ZooKeeper connections — host, session age, expiry, API version. Empty on a non-replicated single node.

  View source on GitHub