ClickHouse Managed Postgres is GA: pricing and the Layerbase alternative

ClickHouse Managed Postgres became generally available on November 9, 2026, the same day ClickHouse announced its acquisition of Xata. This post previously described beta pricing. That comparison needed replacing, not a footnote.

The product distinction still matters. ClickHouse offers Postgres with native integration into its analytics platform. Layerbase offers Postgres and ClickHouse as separate managed databases on a fixed monthly plan. If your main requirement is managed replication between the engines, that is a reason to choose ClickHouse, not a minor feature to overlook in a price table.

Current pricing, checked November 12

ClickHouse's Managed Postgres pricing documentation prices the full VM configuration, including CPU, RAM, and local NVMe storage. The quote depends on tier, region, architecture, instance configuration, and standbys. Use its calculator for the deployment you intend to run.

For AWS, its GA terms apply from November 9:

• Backup storage equal to provisioned disk is included. Extra backup storage uses the regional ClickHouse Cloud storage rate; the published example is $25.30/TB-month in us-east-1.
• Data transfer uses ClickHouse Cloud transfer rates.
• ClickPipes and WalShadow replication from Managed Postgres into ClickHouse is free.

GCP remains a free private preview with later pricing still being finalized. That is separate from AWS GA pricing.

The old roughly $32/month example and 50% beta discount are no longer a usable current quote. I have removed both from the comparison. We should not infer a new price by doubling an old discounted number either.

For a budget, save the configuration alongside its quote. Include the receiving ClickHouse service if you need analytics, and inspect the transfer rules for where your application and database run. Free replication does not mean the entire analytics deployment is free.

What the integration buys you

Managing two databases is not the same as managing the flow of changes between them. A production pipeline has to handle its initial copy, ongoing updates, interruptions, and lag. That operational work belongs in the comparison along with the monthly bill.

ClickHouse also offers one or two standby replicas. Layerbase's Postgres is a single instance. Backups and point-in-time restore help recover data; they are not a substitute for standby failover.

Branching needs a little precision after the Xata news. ClickHouse already documents PITR-based branches, which create independent deployments from a chosen point in time. Xata's instant branching and scale-to-zero integration are announced additions, not evidence that every Xata workflow is available there today. See what Xata users should do now for that transition.

What Layerbase costs

Layerbase's plans, checked November 12, are Free at $0, Solo at $5/mo, and Pro at $15/mo.

Postgres is available on Free, with two databases, 5 GB of storage, and idle sleep. Solo adds capacity to keep one database always-on. ClickHouse requires Pro and is always-on. Pro includes up to ten databases sharing a 1.5 GB RAM / 1 vCPU pool and 25 GB of storage. Pro can add $10/mo pool blocks for more capacity.

Those resources are shared across your databases. A small Postgres and ClickHouse can fit in the included Pro pool, but this is not equivalent to buying a dedicated 8 GB Postgres VM. Size the workload before deciding a cheaper plan is suitable.

Layerbase does not charge per query or per GB transferred. Shared plans do have bandwidth fair-use allowances, so the service is not unlimited. Pro's allowances are 100 GB per day, 500 GB per week, and 1,500 GB per month. Check all three windows for a backfill or analytics export.

Side by side

When I would choose each

Choose ClickHouse Managed Postgres when the integrated analytics path is part of the requirement. Price its standby configuration if failover matters to your application. The service's value is not limited to how cheaply you can keep a Postgres process running.

Consider Layerbase when you need a small managed database, or a few engines with predictable subscription costs, and can accept the resource limits and single-instance design. You get standard endpoints, but two engines still mean two connection strings and your own data-loading or replication code.

For example, a dashboard that can load yesterday's aggregate data in a scheduled application job may not need continuous CDC. A customer-facing view that must reflect transactions almost immediately is a different workload. Start with the freshness and recovery requirements before choosing either plan.

FAQ

Is the beta discount still the basis for comparison?

No. Use the current calculator and GA billing terms. The beta-era price in the previous version of this article has been removed rather than presented as a current offer.

Does running both engines on Layerbase include synchronization?

No. You provision separate databases. Your application or a separately operated pipeline moves data between them.

Does Pro give every database its own resource pool?

No. Pro's included capacity is shared. More databases do not multiply the included RAM, CPU, or storage. Add capacity or evaluate a dedicated server when the workload needs it.

Should Xata users migrate immediately?

Xata has announced no deadline and says customers can continue normally until feature parity enables its managed transition. Moving is an option to evaluate, not an obligation created by the announcement.

Try the workload before moving it

Create a free Postgres database to test the application connection and schema. If the comparison needs analytics too, ClickHouse is available on Pro. Test representative queries and the data-loading path, then compare resource usage and recovery requirements with the quoted ClickHouse deployment.