Editor’s note: This Altinity review is based on my hands-on evaluation of Altinity.Cloud (Cloud Manager 26.4.35 with Altinity Stable 26.3 builds) and Altinity’s open-source tooling in October 2026. I weighed pricing, setup, operations, Antalya, open-source tooling, support, security and rivals.
Quick verdict: I recommend Altinity for engineering teams that want open-source ClickHouse run by specialists, especially inside their own AWS, GCP, Azure or Hetzner account, with the option to walk away on the same open-source stack. The trade-off is per-server sizing and billing by region, and response-time SLAs without a published uptime percentage.
Altinity.Cloud runs ClickHouse in Altinity’s cloud or as BYOC in yours, and Altinity also maintains the Apache 2.0 ClickHouse operator and Stable builds self-hosters run free. In this review, I’ll take a closer look at Altinity’s pricing, features, and support, so you can see exactly whether its BYOC flexibility beats ClickHouse Cloud’s serverless convenience.
Key Takeaways:
- BYOC starts at $0.0625 per vCPU-hour (about $45 per vCPU-month), with infrastructure on your own cloud bill.
- Hosted on AWS US East starts at $0.139 per server-hour plus $0.1056 per vCPU-hour, with 24/7 support included.
- The self-serve 14-day trial includes 16 vCPUs and 1 TB.
- Altinity publishes no uptime percentage, and SSO is set up through a support ticket.
- Antalya adds Iceberg queries and swarm clusters, but reads Iceberg V2 only.
Pros and Cons
Pros
- Runs open-source ClickHouse, with Altinity Stable builds supported for 3 years
- BYOC from $0.0625 per vCPU-hour, with storage left on your own cloud bill
- Hosted on AWS, GCP, Azure or Hetzner, with Hetzner at $0.05 per vCPU-hour
- 24/7 support included on hosted plans; Premium BYOC has a 1-hour response SLA
- Self-serve 14-day trial with 16 vCPUs and 1 TB of storage
- Apache 2.0 Kubernetes operator (2,576 GitHub stars) if you self-host
Cons
- No published uptime percentage, only response-time SLAs
- Hosted bills stack a $0.139 server fee, vCPU rates and storage that vary by region
- SSO and swarm clusters are switched on through a support ticket
- Antalya reads Iceberg V2 only; writes still need Spark or pyiceberg
- Self-managed support contracts are quote-only
How Much Does Altinity Cost?
Altinity sells the same ClickHouse service two ways, both billed hourly: hosted, where a server fee, a vCPU rate and storage cover infrastructure plus 24/7 support, and BYOC, where a flat per-vCPU management fee sits on top of your own cloud bill.
- Altinity.Cloud hosted ($0.139 per server-hour + vCPU + storage): For one invoice covering infrastructure and management. AWS US East adds $0.1056 per vCPU-hour and $0.176 per GB-month; Hetzner’s 6 regions add $0.05 per vCPU-hour and $0.106 to $0.11 per GB-month. 24/7 Elite Support, design assistance, Data Lake Integration and up to $500 a month of network costs per environment are included.
- BYOC Dev ($0.0625 per vCPU-hour, about $45 per vCPU-month): For non-critical workloads that can live with next-business-day support.
- BYOC Enterprise ($0.0833 per vCPU-hour, about $60): For production, with a 4-hour response SLA and design assistance.
- BYOC Premium ($0.1111 per vCPU-hour, about $80): For teams that need a 1-hour response SLA.
| Option | Compute rate | Storage | Support response | Who pays the infrastructure |
|---|---|---|---|---|
| Hosted on AWS US East | $0.139/server-hour + $0.1056/vCPU-hour | $0.176/GB-month | 24/7 Elite Support | Included in Altinity charges |
| Hosted on Hetzner | $0.139/server-hour + $0.05/vCPU-hour | $0.106 to $0.11/GB-month | 24/7 Elite Support | Included in Altinity charges |
| BYOC Dev | $0.0625/vCPU-hour | Your cloud bill | Next business day | You, separately |
| BYOC Enterprise | $0.0833/vCPU-hour | Your cloud bill | 4 hours | You, separately |
| BYOC Premium | $0.1111/vCPU-hour | Your cloud bill | 1 hour | You, separately |
| Self-managed Enterprise Support | Quote-only support contract | Your infrastructure | By severity, 8×5 to 24/7 coverage | You, separately |
Hosted rates move with region: AWS runs from $0.1056 to $0.1707 per vCPU-hour and $0.176 to $0.3344 per GB-month across 34 regions. On BYOC, the management fee covers compute only, so your cloud provider bills the disks.
One 4 vCPU hosted server in AWS US East costs 0.139 + 4 x 0.1056 = about $0.56 an hour, or about $400 a month at 720 hours before storage. A two-replica production cluster starts near $810 a month plus storage, while the same 8 vCPUs on BYOC Dev cost about $360 a month in Altinity fees on top of your AWS bill.
- Prepay: Up to 30% off the whole bill for a prepaid commitment.
- Marketplace billing: AWS and GCP Marketplace spend can qualify for private pricing, and cloud credits can pay for Altinity.Cloud.
- Activity schedules: Stopped clusters stop accruing support charges, but storage keeps billing.
ClickHouse Cloud Basic charges $0.2181 per 2 vCPU/8 GiB unit-hour (about $0.11 per vCPU-hour) and $25.30 per TB-month for storage. Compute lands close to Altinity’s AWS vCPU rate, but storage works out to about $0.025 per GB, roughly a seventh of Altinity’s $0.176 hosted block storage.
Is Altinity Good Value for Money?
- Good fit: BYOC Enterprise at $0.0833 per vCPU-hour if you hold AWS or GCP commitments and want ClickHouse engineers on a 4-hour SLA without moving data out of your account.
- Less convincing fit: Storage-heavy hosted workloads, where $0.176 per GB-month on AWS US East costs far more than ClickHouse Cloud’s object storage.
- Budget priority: Count servers, not just vCPUs, because the $0.139 server fee applies to every hosted replica.
Reviewer’s Notes: Start on the free trial, then move production to BYOC Enterprise if you already have AWS or GCP commitments or credits. Pick hosted on Hetzner for the lowest bill if its European regions or two US datacenters suit you, and pay for Premium only when a 1-hour response changes your incident plan.
My Experience With Altinity.Cloud
Signing Up for the Trial
I signed up at acm.altinity.cloud/signup, which takes a name and email or a Google or Okta login. Email signups need validation and a password of at least 12 characters, and no approval step stands between signup and the console.
The trial runs 14 days with 16 vCPUs and 1 TB of storage, enough to try a two-replica Production layout rather than a single development node. When it expires, clusters keep running but cannot be managed in ACM and DELETE operations are blocked, so ask support for an extension before day 14 if you need time to finish or clean up.
Converting to a billable account means choosing Enterprise or Premium support and paying by credit card. Quotas then rise 10x to 160 vCPUs and 10 TB.
Launching My First Cluster
Environment Setup asks whether the cluster runs in Altinity’s cloud or in your own account. The Launch Cluster Wizard’s Quick Path then offers two presets:
- Development: 1 replica, stopped automatically after 24 hours of inactivity.
- Production: 2 replicas, always on.
The cluster name and availability zones cannot change after launch, so I settled both before moving on. Altinity Builds are the default version, with three years of support, and upstream ClickHouse builds are the alternative. The admin user is fixed as admin, and its 12+ character password appears only once, so copy it before closing the wizard.
The HA tab sets 1 to 3 replicas and a dedicated or shared ZooKeeper/Keeper cluster. Altinity recommends shards only above about 5 TB of data, so most first clusters need replicas, not shards. The advanced settings also hold the backup schedule, IP restrictions and activity schedule, so a cluster launches with its cost controls already set.
Connecting and Running Queries
IP restrictions are on by default, so I added my address before connecting. Traffic enters through the Altinity Edge Ingress load balancer on three ports:
| Connection | Port | Availability |
|---|---|---|
| Native ClickHouse over TLS | 9440 | Hosted and BYOC |
| HTTPS | 8443 | Hosted and BYOC |
| MySQL protocol | 9004 | BYOC only |
The SQL Console splits its tools into tabs:
- Query: Runs SQL.
- Schema: Browses databases and tables.
- Workload: Inspects the query workload.
- DBA Tools: Administrative functions.
- API Endpoints: A beta feature, not yet a substitute for Tinybird’s API layer.
ClickHouse Cloud is serverless and autoscales, while Altinity asks you to choose node sizes. That means more work at launch and a more predictable bill afterward.
Reviewer’s Notes: Pick the smallest node type that handles the workload, and add a replica before a shard. Switching an 8 vCPU node to a 4 vCPU node with the same RAM can nearly halve the compute part of the bill, though the $0.139 server fee and storage stay the same.
Running ClickHouse Day to Day: Scaling, Backups and Monitoring
ACM puts upgrade, stop, resume and restart behind buttons on each cluster, alongside backup, restore, clone and replica-cluster actions. Each environment runs on a dedicated Kubernetes cluster that Altinity manages, except in bring-your-own-Kubernetes setups. Activity schedules decide what idle clusters cost:
- Always On: For production clusters.
- Stop When Inactive: Shuts idle clusters down, the right setting for development.
- Stop On Schedule: Stops clusters at set times, such as outside working hours.
- Rescale On Schedule: Changes node size on a timetable for batch windows.
- Rescale When Inactive: Shrinks a cluster when activity drops.
Backups run on schedules from hourly to monthly, up to three per cluster, with a default retention you can change. AWS backups older than 15 days move to Infrequent Access storage. On Enterprise Support, urgent restore requests get a response in under 4 hours, which is a response time, not a restore time.
Monitoring runs on Prometheus, Loki and Grafana, and ACM surfaces the alerts, logs, dashboards and health checks. A separate grafanauser role gives on-call engineers dashboard access as its own permission.
For automation, the ACM REST API at acm.altinity.cloud/api takes an X-Auth-Token header, with keys created under My Account > API Access. Keys expire after 24 hours by default, so unattended jobs need a longer expiry or a refresh step. The altinity/altinitycloud Terraform provider uses the Anywhere API token to rebuild environments from code instead of repeating wizard settings.
Aiven puts ClickHouse, Kafka and PostgreSQL in one console, which suits teams that want a single vendor for several data services.
Best for: ClickHouse-focused teams that want explicit control of node size, schedules and backups. Skip if: you would rather hand capacity decisions to an autoscaler.
Project Antalya: Querying Iceberg Data Lakes With Swarm Clusters
Project Antalya builds are open-source ClickHouse builds under Apache 2.0 that add an Iceberg database engine over REST catalogs, AWS Glue and S3 Table Buckets. They also bring:
- Parquet improvements: Better Parquet reads and writes against object storage.
- Swarm clusters: Stateless query workers that take the heavy reads.
- Hybrid tables: MergeTree and Iceberg data behind one table.
Altinity also maintains Ice, an Apache 2.0 toolset for running Iceberg REST catalogs, so the catalog side does not tie you to Altinity.Cloud. Data Lake Integration carries no extra charge on hosted or BYOC, and Altinity publishes no separate Antalya price.
Antalya support is measured in months, not years: build 26.6.8.20002 (October 15, 2026) is supported to February 26, 2027, and 26.3.53.20001 (October 21, 2026) to January 3, 2027. The older 25.8.33.20002 build loses support on November 5, 2026. A Stable build gets three years, so Antalya means more frequent upgrades.
The limitations decide whether a pilot goes further:
- Iceberg V2 reads only: Writes still need Spark or pyiceberg.
- Roadmap gaps: Tiered storage moves (ALTER TABLE MOVE, TTL MOVE) and materialized views on Iceberg are not shipped yet.
- Version lock: Every node in a distributed query must run the same Antalya version; mixed versions during a rolling upgrade may fail or return wrong results.
- Support-enabled swarms: Altinity support has to switch swarm clusters on.
StarRocks and Apache Doris already query Iceberg through lakehouse catalogs, and they are the stronger choice if Iceberg is your primary store.
Pilot Antalya if your cold data already sits in Iceberg and you want to query it with ClickHouse SQL next to MergeTree tables. It is not yet a replacement for a lakehouse engine that has to write Iceberg.
Open-Source Tooling and Support for Self-Hosted ClickHouse
The Altinity Kubernetes Operator for ClickHouse is Apache 2.0, has 2,576 GitHub stars, and shipped 0.27.4 on September 24, 2026, roughly one release every one to two months. It has been FIPS compatible since 0.27.1. The rest of the self-hosting stack:
- Altinity Stable builds: 26.3.53.10001 shipped July 3, 2026, with extended support to July 3, 2029. Every Stable build gets 3 years, and FIPS 140-3 builds exist.
- clickhouse-backup: v2.8.2 shipped October 21, 2026; the project has 1,654 GitHub stars.
- Enterprise Support: Paid help for self-managed ClickHouse, with SLAs from 8×5 to 24/7 and response times by severity, priced on a call only.
Extended support covers P0 and P1 bugs and critical security issues on a best-effort basis, so treat it as a maintenance window rather than a fix guarantee.
GitHub users have reported Keeper and upgrade friction on older operator releases:
- 0.25.6: Keeper installs stuck in progress with no clear error.
- 0.24: An upgrade path that required deleting and recreating Keeper.
- 0.25.5: Readiness probes that broke fresh Keeper installs.
- 0.25.0 and 0.25.1: Crash loops after upgrading.
- Helm chart 0.25.6: A CRD ConfigMap that exceeded the 262,144-byte limit.
The operator repo carries 161 open issues. None of the reports above concern 0.27.4, but test Keeper creation and operator upgrades in staging before touching production.
On BYOC, Altinity says you can disconnect its control plane and keep running the open-source components. Instaclustr also offers BYOC ClickHouse but publishes no prices, so Altinity’s per-vCPU BYOC fees are easier to budget before a sales call.
I recommend Enterprise Support for teams already running the operator who want an escalation path, and Altinity.Cloud for teams that do not want to own Kubernetes upgrades.
Security, Access Control and Compliance
Altinity.Cloud is SOC 2 Type II certified, and Altinity names no HIPAA, PCI DSS or ISO 27001 certification, so check mandatory frameworks before procurement.
- Isolation: Tenant environments separated in compute, storage and network.
- Encryption: Data at rest and backups encrypted, with TLS endpoints.
- Availability: Replicas across availability zones.
- Roles: orgadmin, envadmin, envuser, envsupport, grafanauser and billing, scoped over an organization > environment > cluster > database hierarchy.
- Cloud access: Temporary cloud credentials for the environments Altinity operates.
- Console logins: Google Auth, with support handled through Zendesk and Slack.
SSO runs through Auth0 with Google, Microsoft Entra, Okta or Keycloak, but it is not self-serve: you send your domain, client ID and secret to Altinity support. Agree on a secure channel for that handover before onboarding production admins.
Aiven for ClickHouse covers SOC 2, ISO 27001, HIPAA, GDPR and PCI DSS with a 99.99% SLA, so regulated teams may find its paperwork easier to clear. Altinity backs its response-time SLAs with performance penalties, but a response SLA does not answer an uptime question.
How Does Altinity Compare to Competitors?
Altinity wins on open-source portability and BYOC, while ClickHouse Cloud wins on serverless autoscaling and cheaper object storage.
| Tool | Starting price | Deployment | Best at |
|---|---|---|---|
| Altinity.Cloud | Hosted AWS US East: $0.139/server-hour + $0.1056/vCPU-hour, storage extra; BYOC from $0.0625/vCPU-hour; 14-day trial | Hosted or BYOC on AWS, GCP, Azure, Hetzner; any Kubernetes | Open-source ClickHouse with specialist support |
| ClickHouse Cloud | Basic: $0.2181 per 2 vCPU/8 GiB unit-hour; $300 trial credits for 30 days | Serverless SaaS | Autoscaling and inexpensive object storage |
| Aiven for ClickHouse | Usage-based plans | Multi-cloud SaaS | ClickHouse, Kafka and PostgreSQL together; 99.99% SLA |
| Instaclustr | Quote-only | SaaS or BYOC | Managed ClickHouse with up to 99.99% SLA |
| Tinybird | Free tier; Developer $49/month | SaaS | APIs built on ClickHouse |
| StarRocks | Free open source | Self-hosted or vendor cloud | Lakehouse joins |
- Altinity vs ClickHouse Cloud: ClickHouse Cloud suits teams that want no node-sizing decisions, and its $25.30 per TB-month storage pays off on large datasets. It does bill internet egress at $0.1152 per GB, while Altinity’s hosted plans include up to $500 a month of network costs per environment. Altinity makes more sense when running the managed service inside your own cloud account outweighs the storage savings.
- Altinity vs Aiven: Aiven suits teams already running Kafka and PostgreSQL next to ClickHouse in one console, and its HIPAA and PCI DSS coverage plus a 99.99% SLA help regulated procurement. Altinity stays focused on ClickHouse operations and its open-source tooling.
- Altinity vs Instaclustr: Compare Instaclustr when BYOC and an uptime commitment are both mandatory, since it advertises up to 99.99% availability. Altinity counters with published per-vCPU BYOC fees you can budget before any sales call.
- Altinity vs Tinybird: Tinybird suits developers shipping analytics as APIs rather than running a cluster; its free tier includes 10 GB and 1,000 requests a day, and Developer costs $49 a month. Altinity fits when cluster configuration is part of the job.
- Altinity vs StarRocks: StarRocks is the stronger pick for lakehouse joins. Antalya keeps ClickHouse SQL and MergeTree next to Iceberg, but its read-only Iceberg support and version lock call for a pilot first.
DoubleCloud shut down on March 1, 2025, so former users comparing managed ClickHouse can drop it from the list.
Compare deployment model before headline compute price: Altinity’s BYOC fee excludes infrastructure, while hosted rivals bundle it, so setting a management-only rate against a full compute charge produces the wrong budget. Likewise, a 1-hour Premium response answers a different contractual question than an Aiven or Instaclustr uptime SLA.
How I Reviewed Altinity
I worked through the Altinity.Cloud trial flow, the Launch Cluster Wizard, connection options and the SQL Console in Cloud Manager 26.4.35, then the scaling schedules, backup settings, monitoring, API and Terraform provider. I also covered Project Antalya’s Iceberg and swarm features, the open-source operator and Stable 26.3 builds, and the security roles and SSO setup.
For cost, I compared hosted and BYOC rates on AWS, GCP, Azure and Hetzner, keeping Altinity’s management fees separate from infrastructure costs. I compared Altinity with ClickHouse Cloud, Aiven, Instaclustr, Tinybird and StarRocks on price, deployment model and support terms, and I make no benchmark or recovery-time claims.
Prices current as of October 2026.
Should You Run ClickHouse With Altinity?
Altinity is the right way to run open-source ClickHouse with specialist support, especially when the cluster has to live in your own cloud account. Its strengths are BYOC flexibility and Stable’s 3-year support window, not serverless simplicity or the cheapest storage.
- Cloud commitments or residency rules: Choose BYOC Enterprise for its 4-hour response SLA and design assistance while keeping data in your environment.
- Cost-focused teams in Europe: Choose hosted on Hetzner, at $0.05 per vCPU-hour, if its regions fit.
- Teams already self-hosting with the operator: Buy Enterprise Support for an escalation path without replacing your deployment.
- Hands-off serverless workloads: Choose ClickHouse Cloud for autoscaling and cheaper object storage.
- HIPAA or PCI procurement: Start with Aiven’s broader compliance coverage.
For most buyers, BYOC Enterprise is the production fit; Premium earns its extra $0.0278 per vCPU-hour only when a 1-hour response changes your incident plan.
Start the 14-day trial at Altinity and launch a Development cluster with Altinity Builds.
FAQs
How much does Altinity.Cloud cost?
Hosted Altinity.Cloud costs $0.139 per server-hour plus a regional vCPU rate and storage. On AWS US East that is $0.1056 per vCPU-hour and $0.176 per GB-month, with 24/7 support included. BYOC costs $0.0625 to $0.1111 per vCPU-hour in management fees, and your cloud provider bills the infrastructure.
Is Altinity the same as ClickHouse Cloud?
No. Altinity is a separate company that runs open-source ClickHouse and its own Altinity Stable builds in Altinity’s cloud or as BYOC in yours. ClickHouse Cloud is the serverless service from the ClickHouse company, with autoscaling and object storage at $25.30 per TB-month. Pick Altinity for BYOC and specialist support, ClickHouse Cloud for fewer sizing decisions.
Does Altinity have a free trial?
Yes, a 14-day self-serve trial with 16 vCPUs and 1 TB of storage. You can sign up with an email address, Google or Okta, and support can extend the trial on request. Plan for expiry: clusters keep running, but you cannot manage them in ACM and DELETE operations are blocked.
Is the Altinity ClickHouse operator free?
Yes. The Altinity Kubernetes Operator for ClickHouse is Apache 2.0 licensed, with 2,576 GitHub stars and release 0.27.4 from September 2026. Support is the paid part: Enterprise Support for self-managed ClickHouse is quote-only, with SLAs from 8×5 to 24/7. The license costs nothing, but upgrades, backups and incidents stay your job.
Can Altinity run in my own AWS account?
Yes. Altinity.Cloud runs as BYOC in your own AWS, GCP, Azure or Hetzner account, and as bring-your-own-Kubernetes on any cluster. You pay Altinity $0.0625 to $0.1111 per vCPU-hour for management, while compute and storage stay on your cloud bill. AWS and GCP Marketplace billing can qualify for private pricing.


