ClickHouse support, 24/7

ClickHouse Support for Self-Managed Clusters, Keeper, Kubernetes and ClickHouse Cloud

Last reviewed by Tyler Eastridge

AceMQ supports ClickHouse 24/7 on self-managed clusters, ClickHouse Keeper, Kubernetes and ClickHouse Cloud. Most calls start with one of three problems: "too many parts" errors from small inserts, queries killed by memory limits on a join or GROUP BY, or a replication queue that stopped draining behind one failing entry.

24/7 across Americas, EMEA and APAC, with response targets from 15 minutes, by plan, and a written root cause analysis on every P1.

Support quotes answered within 1 hour, by a senior engineer.

Planning a build, upgrade or migration instead? See AceMQ ClickHouse consulting and services.

160+
Customers supported
10+
Years in production support
40+
Sev 1 incidents handled a year
24/7
Americas, EMEA and APAC
1hour
Quote reply

Customers

  • Microsoft
  • Henry Schein
  • Enbridge
  • Gojek

Partners

  • Broadcom
  • Redis
  • VMware Tanzu
  • OSSeva

Incident Triage

ClickHouse problems we fix every week

Part counts, memory limits, sparse-index misses and replication queues: eight ClickHouse symptoms and the system table we query first for each.

Common ClickHouse production problems and how AceMQ resolves them
Symptom you're seeingWhat we check firstTypical resolution
DB::Exception: Too many parts on insertsystem.parts active part count per partition and the insert rate; system.part_log for merge lag; partition key cardinalityBatch at the producer, or enable async_insert; partition by day, not by timestamp or user id; raising the parts limit is a stopgap
Memory limit exceeded on a JOIN or GROUP BYWhich table is on the right side of the hash join; max_bytes_before_external_group_by; system.query_log memory per queryReorder the join, pick the join algorithm deliberately, let aggregation spill to disk
A filtered query scans the whole tableWhether the WHERE column is a prefix of the ORDER BY key; granules read versus granules matching in EXPLAIN indexes = 1A projection, a skipping index or a corrected sorting key
Replicas diverge, system.replication_queue keeps growingThe oldest queue entry with a growing num_tries; Keeper latency and session expiriesClear the blocking entry deliberately; size Keeper nodes
Disk fills despite TTL rulessystem.parts for unmerged parts, system.mutations for a long rewrite, free space against what a merge needsFree space first, then let merges and TTL run
An ALTER TABLE UPDATE or DELETE runs for dayssystem.mutations parts_to_do; mutations rewrite every affected partLightweight delete, a ReplacingMergeTree pattern, or a partition-scoped rewrite
Distributed queries are far slower than on one shardShard skew from the sharding key, and whether the initiator re-aggregates a large intermediate resultPush the GROUP BY down and return aggregate states
Ingest slowed after adding a materialized viewNumber of views fired per insert block and each target's merge loadDrop or consolidate views; each runs in the insert path

A P1 closes with a written root cause naming the table, key or query pattern responsible and the schema or setting change that removes it.

Inserts failing with too many parts, or a replica that will not catch up? Request a quote and a ClickHouse engineer replies within an hour.

Contact Us

Deployment coverage

Which ClickHouse deployments does AceMQ support, and who covers what?

DeploymentWhat AceMQ coversWhat the vendor coversExcluded
Self-managed ClickHouse on any release, VMs or bare metal24/7 incident response, written RCA on P1s, schema and sorting-key review, ingest tuning, Keeper health, upgradesClickHouse publishes security fixes for a short list of versions (SECURITY.md, read 17 Nov 2026); the community answers questions on SlackApplication code is reviewed, not rewritten
ClickHouse Cloud (AWS, GCP, Azure)The workload layer: schema, sorting keys, join memory, mutation cost, insert batching, query tuningClickHouse runs the service and provides support for Cloud customers with 24x7 coverage for Sev-1 (clickhouse.com/support/program, read 17 Nov 2026)Platform and service limits set by ClickHouse
ClickHouse Keeper and ZooKeeper clustersSizing, latency, session handling, ensemble recoveryApache and ClickHouse projects for their softwareHardware and network faults
Altinity Stable builds and the Altinity OperatorThe ClickHouse behaviour on those builds and the operator's effect on pods, volumes and rolling restartsAltinity supports its builds and operatorAltinity-only components
Kubernetes and OpenShiftPersistence on volumes, resource limits against merge memory, rolling restarts, shard and replica placementThe Kubernetes vendor supports the platformControl-plane faults
Tiered storage on S3 or GCSMove rules, cold-read latency, TTL-to-volume behaviourThe cloud provider supports object storageObject store faults

Support options

Who provides enterprise support for ClickHouse? ClickHouse and independent support compared

ClickHouse, Inc. provides support to ClickHouse Cloud customers. Self-managed ClickHouse has no equivalent contract on the support page we read, so enterprise support for it comes from independent providers or your own team.

ClickHouse, Inc. support
AceMQ independent support
Who is coveredUsers and customers of ClickHouse Cloud; the page describes only Cloud (clickhouse.com/support/program, read 17 Nov 2026)
Who is coveredSelf-managed, Kubernetes and Cloud workloads on any release
Hours24x7 for Sev-1 and business days for Sev-2 and Sev-3 on the Cloud plan
Hours24/7 across Americas, EMEA and APAC
Initial response, Sev-1Scale 1 hour, Enterprise 30 minutes, both 24x7
Initial response, Sev-1From 15 minutes, by plan
Initial response, Sev-2Scale 4 business hours, Enterprise 2 business hours
Initial response, Sev-21 hour
Initial response, Sev-31 business day
Initial response, Sev-34 hours for P3; next business day for P4
Product defectsClickHouse can fix and ship its own code
Product defectsWe reproduce, document and work the escalation
Self-managed ClickHouseNot described on the support page
Self-managed ClickHouseCovered
PricePer ClickHouse Cloud plan
Price$24,000 to $72,000 a year (USD), set by SLA, scope and deployment. Quote reply within 1 hour

Short note: ClickHouse support covers the Cloud product. Most incidents on any ClickHouse are operational (insert shape, sorting keys, mutations, Keeper), which is why teams that run Cloud also hold independent cover for the workload.

What's Included

Everything in your ClickHouse support contract

MergeTree schemas, Keeper, replicas, shards, mutations and materialized views are handled as one system, not as separate tickets.

  • 01

    Emergency Incident Response

    Ingest stopped, queries failing, replicas diverged, or a disk filling faster than merges can free it. With response targets from 15 minutes, by plan, a senior engineer joins a live bridge with access to diagnose — not a ticket acknowledgement.

  • 02

    Root Cause Analysis

    Every P1 closes with a written RCA: what failed, why, the fix applied, and the schema, ingest, or settings change that stops it recurring. Delivered as standard, not on request.

  • 03

    Schema & Table Engine Review

    Sorting key and primary index design against your real filter patterns, partition granularity, codec and compression selection per column, and choosing between ReplacingMergeTree, AggregatingMergeTree, and plain MergeTree deliberately.

  • 04

    Query & Ingest Tuning

    EXPLAIN and query log analysis, join reordering and algorithm selection, projections and skip indexes, plus insert batching or async_insert configuration sized against your producer's actual behaviour.

  • 05

    Replication & Keeper Health

    ReplicatedMergeTree topology, ClickHouse Keeper sizing and latency, replication queue monitoring, and distributed table design that doesn't create hot shards or expensive fan-out.

  • 06

    Upgrades & Migrations

    Version upgrades with a tested rollback, self-managed to ClickHouse Cloud moves in either direction, tiered S3 storage rollout, and migrations onto ClickHouse from Elasticsearch, Druid, or a row-store warehouse.

Why AceMQ

Why analytics teams bring AceMQ onto their ClickHouse estate

  • Named Engineers, Zero Cold Start

    The same senior engineers stay on your account. They know your table engines, your sorting keys, and your ingest pattern — so a P1 call starts with a hypothesis rather than a walkthrough of your schema.

  • No Tier-1 Triage Layer

    You reach a senior ClickHouse engineer directly by phone, email, or Slack. Nobody collects information to pass along, and no escalation approval stands between you and someone who can read system.query_log.

  • We Fix the Schema, Not Just the Settings

    Most ClickHouse incidents trace back to a sorting key that doesn't match the query filters or a partition key with far too much cardinality. Settings changes can buy time; we tell you when the real fix is a table rebuild and plan the migration.

  • Ingest Pipelines, Not Just the Server

    "Too many parts" is a producer problem that manifests in the database. We tune the Kafka table engine, batching in your ingest service, and async_insert together, because fixing only the server side just moves the failure later.

  • Genuine Follow-the-Sun Coverage

    Round-the-clock coverage for customers in 26+ countries. Your 3am ingest stall is someone's mid-afternoon — no overnight skeleton crew, no waiting for a region to come online.

  • Full-Stack, Not Just clickhouse-server

    We diagnose across Keeper, disk and object-store latency, the ingest pipeline, network policy, and Kubernetes. ClickHouse is fast enough that the bottleneck is frequently somewhere else entirely.

ClickHouse layouts we run support for

Self-managed clusters on VMs or Kubernetes, replicated layouts coordinated by Keeper, sharded deployments, and ClickHouse Cloud where schema and query design are still yours.

  • ClickHouse Cloud (AWS, GCP, Azure)
  • Self-managed ClickHouse on any release; ClickHouse's SECURITY.md lists 26.10, 26.9, 26.8 and 26.3 for security fixes (read 4 Nov 2026)
  • ClickHouse Keeper & ZooKeeper clusters
  • Altinity Stable builds
  • Altinity Operator on Kubernetes
  • OpenShift
  • AWS (EC2, EKS) with gp3 / io2 volumes
  • Azure & Google Cloud VMs
  • S3 / GCS tiered storage
  • Bare metal & on-premise
  • Air-gapped / no outbound access
  • Hybrid cloud

Version coverage

ClickHouse versions we support

ClickHouse ships frequent releases. AceMQ supports the release you run and plans the move to a version that receives security fixes.

ClickHouse versions we support
ClickHouse releaseListed for security fixes (SECURITY.md)What AceMQ does on it
26.10Yes24/7 support
26.9Yes24/7 support
26.8Yes24/7 support
26.3Yes24/7 support
26.7 to 26.1, 25.x and earlierNot listed24/7 support and an upgrade plan to a listed release
ClickHouse CloudManaged by ClickHouse24/7 support for the workload layer

Source: github.com/ClickHouse/ClickHouse/blob/master/SECURITY.md, read 17 November 2026. Upgrades go through a staging rehearsal against your real query mix, replicas first, with a tested rollback.

In practice

Why most ClickHouse incidents are schema incidents

ClickHouse is fast enough that the bottleneck is usually the shape of the data. Three design decisions decide most of the tickets we see.

Sorting key. The primary index is sparse over the ORDER BY key and prunes granules only when the filter matches a key prefix. A key that leads with a high-cardinality identifier gives time-range dashboards almost no pruning.

Partition key. Each insert creates at least one part per partition it touches. A high-cardinality partition key multiplies parts and starves merges.

Mutation habits. ALTER ... UPDATE and DELETE rewrite every affected part. Routine row corrections done this way rewrite the whole dataset.

We tell you when the fix is a table rebuild and plan the migration so the old table stays queryable until the new one is verified.

Consulting

ClickHouse consulting and professional services

ClickHouse consulting and professional services
DeliverableWhat you receive
Cluster and storage assessmentProfile under sustained load to find which resource binds first (merges, concurrency, Keeper, storage bandwidth), a capacity model tied to projected volume
Schema and sort key designGranules read versus granules matching per query, reordered sort keys, partition granularity, per-column codecs, materialised views and aggregate tables, TTL policy
Ingest reviewBatch size and async_insert sizing, Kafka table engine tuning, materialized view fan-out
Replication and Keeper reviewReplicatedMergeTree topology, Keeper sizing and latency, runbooks for stalled queue entries
Upgrade planVersion and client inventory, staging rehearsal, rolling order, rollback
Migration planningMoves onto ClickHouse from Elasticsearch, Druid or a row-store warehouse, and between self-managed and Cloud; see ClickHouse migration services

Escalation Path

Your first hour of a ClickHouse outage

The engineer who answers already knows what system.parts and system.replication_queue should look like, so we begin by reading them.

  1. From 15 min (by plan)first responseEmergencyProduction down: inserts rejected for too many parts, queries failing on every shard, or a full disk blocking merges
  2. 1 hourfirst responseCriticalSevere impact: replicas diverging with a growing replication queue, memory limit errors on joins, Keeper sessions expiring
  3. 4 hoursfirst responseHighMutations running for days, materialized views slowing ingest, TTL rules not freeing space, skewed shards
  4. Next Dayfirst responseStandardQuestions on ORDER BY design, skipping indexes, async_insert, join order, upgrade rehearsal and Cloud versus self-managed

Ingest stopping on a live dashboard and a question about a sorting key are on separate clocks, chosen by plan.

Prefer a named engineer who already knows your tables and sorting keys? Request a quote or message us.

Contact Us

Pricing

What does AceMQ ClickHouse support cost?

ClickHouse support from AceMQ usually costs $24,000 to $72,000 a year. It scales with your shard and replica layout, data volume and ingest rate, whether you run self-managed or Cloud, and the response plan.

Prices are in US dollars (USD). Response targets start at 15 minutes by plan, 24/7, and we answer quote requests within 1 hour. Covering Kafka, Druid or other data tools in the same contract earns a multi-technology discount.

Support In Practice

ClickHouse problems we have unpicked

Typical ClickHouse support, remediation and consulting engagements. Each card links to the full write-up: the cluster, what the engagement covers and what it delivers.

  • Remediation

    ClickHouse Too Many Parts Remediation

    Adtech Measurement Platform

    Resolving ingestion failures where frequent small inserts produced parts faster than background merges could retire them, tripping the parts limit.

    ClickHouseKafkaKubernetesPrometheus
  • Consulting

    ClickHouse Schema and Sort Key Design Consulting

    Industrial IoT Platform Operator

    Redesigning ORDER BY keys, partitioning, codecs, and materialized views so dashboard queries read a small fraction of the data instead of full scans.

    ClickHouseKafkaGrafanaLinux
  • Support

    ClickHouse MergeTree and Replication Support

    Cybersecurity Analytics Provider

    Ongoing support for ReplicatedMergeTree clusters covering replication queue stalls, memory limit failures on large queries, and mutation backlogs.

    ClickHouseClickHouse KeeperKubernetesPrometheus
All use cases

Support options

Four tiers of ClickHouse support.

AceMQ offers ClickHouse support in four tiers, each handing more of the day-to-day work to AceMQ: 24/7 support and incident management for clusters your team runs, proactive monitoring that catches part counts, mutation debt and Keeper latency first, managed ClickHouse on your own infrastructure, and fully managed ClickHouse on AceMQ infrastructure. Self-managed clusters and ClickHouse Cloud are covered, on any release.

  • Tier 1

    ClickHouse 24/7 Support & Incident Management

    Your engineers run ClickHouse, we take the escalations

    Best for: Analytics platform teams that own their clusters and need a specialist for merge lag, replica drift and heavy queries.

  • Tier 2

    ClickHouse 24/7 Support with Proactive Monitoring

    Reviews of parts, merges, keys and Keeper

    Best for: Teams whose tables grow quickly and who want sorting keys, TTLs and merge pressure checked before ingest slows down.

  • Tier 3

    Managed ClickHouse on Your Infrastructure

    We operate your clusters and Keeper nodes

    Best for: Organizations keeping ClickHouse on their own infrastructure that want shard, replica and upgrade work handled for them.

  • Tier 4

    Fully Managed ClickHouse on AceMQ Infrastructure

    We run your ClickHouse shards and replicas

    Best for: Product teams that want fast analytics queries without owning merges, mutations and replication day to day.

FAQ

ClickHouse support questions

Still have questions about ClickHouse? Email an Expert

Who provides enterprise support for ClickHouse?

ClickHouse, Inc. supports its Cloud customers with published response times. For self-managed clusters on VMs or Kubernetes, independent providers like AceMQ cover operations, schema and query problems around the clock.

What does 24/7 ClickHouse support cover with AceMQ?

A senior engineer on a live call at any hour for P1s such as stopped ingest, queries failing on all shards or a full disk blocking merges, plus a written root cause for every P1 and help with tuning and upgrades.

What fixes "too many parts" in ClickHouse?

Batching at the producer, because every insert creates at least one part per partition and merges cannot keep up with thousands of tiny inserts a second. Tens of thousands of rows per insert is normal, or async_insert lets the server buffer. Correct a high-cardinality partition key too. Raising the limit trades an error today for a slower cluster tomorrow.

Why does my filtered query still scan the whole table?

ClickHouse has no per-column indexes like a row store. The primary index is sparse over the ORDER BY key and skips granules only when the filter matches a key prefix. Filtering on the second key column alone prunes nothing. Options are a skipping index, a projection with a different order, or a rebuilt table; we measure which wins on your data.

An ALTER TABLE UPDATE has run for days. Can it be stopped?

Yes, with KILL MUTATION; already rewritten parts stay valid. The better question is why it was needed. Mutations rewrite every affected part. Lightweight deletes, ReplacingMergeTree with FINAL at query time or a partition-scoped rewrite avoid whole-table rewrites.

What causes ClickHouse replication lag and how is it fixed?

Slow or overloaded Keeper, a replica unable to fetch a part from a busy source, or a mutation stuck at the head of the queue. We relieve the coordination service, clear the blocking entry on purpose and size merge pools and storage for the ingest rate. The system tables show it quickly on the first call.

Do you support ClickHouse Cloud as well as self-managed clusters?

Yes, with the same engineers. Cloud removes Keeper and storage provisioning, but schema design, sorting keys, join memory, mutation cost and insert batching still decide most tickets. ClickHouse's own support covers the Cloud product.

Which ClickHouse versions do you support?

The release you run. SECURITY.md currently lists 26.10, 26.9, 26.8 and 26.3 for security fixes (read 17 Nov 2026); we support others and plan the move to a listed release.

Can you upgrade our ClickHouse cluster?

Yes: version and client inventory, a staging rehearsal against your real query mix, replicas first, then a controlled order and a written rollback. We also check materialized views and TTL behaviour on the new version.

What does a ClickHouse health check include?

Part counts and merge lag per table, sorting and partition key review against the query log, memory limit behaviour, replication queue and Keeper latency, mutation backlog, disk and TTL status, and a ranked plan. Quotes are answered within 1 hour.

Who offers ClickHouse consulting and professional services?

ClickHouse, Inc. offers professional services with its Cloud plans, per its support page, and independent firms such as AceMQ consult on self-managed and Cloud ClickHouse alike. Our consultants staff the 24/7 rota, so recommendations come with the knowledge of how a design fails. See ClickHouse consulting.

Can we buy a ClickHouse health check first and decide on support later?

Yes. A health check covers part counts, merge lag, sorting keys, memory limits, the replication queue and mutation backlog, and ends with a ranked plan. Many teams then add annual 24/7 support, often with the same engineer who did the review.

How do you price ClickHouse support?

Annual contracts run $24,000 to $72,000 (USD). The inputs are shard and replica count, data volume and daily ingest, Keeper layout, self-managed or Cloud, and the response plan. Send us those figures and the quote comes back within 1 hour.

Let's talk

Read system.parts Before the Insert Errors Start

Send us your ClickHouse version, table engines and your current part count per partition. We reply to quote requests within 1 hour.

A senior ClickHouse engineer, not a sales queueQuote requests answered within 1 hourAd hoc engagements or an annual partnership

Production incident right now?Email a ClickHouse Architect