Case study 02 ยท Managed ClickHouse
ClickHouseยฎ Cloud priced HA at $892.
The same cluster runs at $335 on Selfhost.dev.An 8-person analytics team serves usage dashboards inside a B2B product: about two billion rows, 400 GB compressed, queried by customers all day. On ClickHouse Cloud, outgrowing the Basic tier's fixed 12 GiB replica meant the Scale tier, two replicas and about $892 a month. The same high-availability shape on Selfhost.dev managed ClickHouse bills about $335, same region. Line items below.
- Team
- 8 people, analytics platform inside a B2B SaaS
- Workload
- Customer-facing dashboards, always-on
- Data
- ~2B events, ~400 GB compressed
- Region
- US, everything in us-east-1
- Trigger
- Forced off the Basic tier by its 12 GiB memory ceiling
Client details are anonymized and normalized. Every price below is recomputed at public list rates and our own billed rates for August 2026, in the region named on each table, so you can check the math yourself.
Why they moved
The ceiling was memory,
the cliff was the tier.
ClickHouse Cloud's Basic tier is a single fixed replica at 8 or 12 GiB. It carried this workload for a year and it was priced fairly for what it was. Then dashboard concurrency grew, the big aggregation queries started touching memory limits and the next rung on the ladder was not a bigger Basic. It was the Scale tier.
Scale is the tier with vertical scaling and high availability and it runs multiple replicas. For a service customers hit all day, the team wanted HA anyway. But the arithmetic stacked fast: two replicas, 16 GiB each, is four compute units on the meter, every hour, at a higher per-unit rate than what they now pay in total.
The before table
What ClickHouse Cloud was charging,
line by line.
ClickHouse Cloud, Scale tier ยท us-east-1 ยท published list rates
| Item | Detail | $ / mo |
|---|---|---|
| Compute | 2 replicas x 16 GiB = 4 units at $0.2985/unit-hr, always-on | $871.62 |
| Storage | 400 GB compressed at $25.30/TB, shared object storage | $10.12 |
| Backups | Default daily backup, billed as storage | $10.12 |
| Total | about $892 every month | $891.86 |
ClickHouse Cloud published list rates, us-east-1, checked August 2026: Scale compute $0.2985 per unit-hour where one unit is 8 GiB RAM + 2 vCPU, storage $25.30/TB/mo compressed, default backup billed as storage. Metering is per minute; this workload is customer-facing and runs 730 hours a month, so it is priced always-on.
The Basic tier is a fixed single replica at 8 or 12 GiB and did not fit this service. Egress and ClickPipes excluded, conservatively, on both sides.
The diagnosis
What was actually
driving the bill.
Basic replicas come in exactly two sizes, 8 or 12 GiB, with no vertical scaling. The moment a workload needs byte thirteen, it needs a new tier, not a bigger box. That is a pricing cliff disguised as a product tier.
On the Scale meter, a unit is 8 GiB welded to 2 vCPU at a fixed 1:4 ratio. Sixteen GiB per replica therefore buys 4 vCPU per replica whether you use it or not. This workload is memory-bound: big aggregations, modest concurrency. It was paying for CPU it never touched.
ClickHouse Cloud's idling is genuinely good economics for bursty or internal workloads. Customer-facing dashboards never idle. Priced flat-out, 730 hours a month, the serverless premium has nothing left to pay for itself with.
The migration
The move,
as it actually went.
Migrations are free and white-glove on Selfhost.dev: an engineer plans and runs the move with you. ClickHouse to ClickHouse is a schema copy, a backfill and a short dual-write window.
Shape-match, without the welded ratio
The memory requirement mapped to two r7g.large nodes, 16 GB each at 2 vCPU, exactly the memory-to-CPU shape the workload wanted. High availability came from a primary plus a same-size replica with an embedded ClickHouse Keeper quorum node. Region stayed us-east-1.
Backfill, then a short dual-write window
Historical partitions were backfilled first, then the ingest pipeline wrote to both sides for a few days while dashboards were checked against each other. Cutover was flipping the read endpoint: HTTP on 8123 and native TCP on 9000, same as before.
Same instincts, fewer meters
Snapshots and backup policies, resize, live metrics and alert rules are all part of the instance and the cluster is manageable from the editor over MCP, which on Selfhost.dev provisions and operates rather than just reading.
The after table
The same HA shape on Selfhost.dev,
line by line.
Selfhost.dev managed ClickHouse ยท us-east-1 ยท all-in billed rates, August 2026
| Item | Detail | $ / mo |
|---|---|---|
| Compute | 2 x r7g.large, 2 vCPU / 16 GB each (primary + replica) | $234.55 |
| Storage | 400 GB gp3 at $0.08/GB base, per node (800 GB total) | $96.00 |
| ClickHouse Keeper | t4g.small + 20 GB, the quorum node for the pair | $20.80 |
| Boot volumes | 20 GB gp3 per node, three nodes, billed at the gp3 rate | $7.20 |
| Public IPv4 | Added at deploy | $5.48 |
| Total | about $364 every month | $364.03 |
Selfhost.dev managed ClickHouse figures are our all-in billed us-east-1 rates as of August 2026. EBS storage is attached per node, so the HA pair carries it twice and every node boots from its own 20 GB gp3 volume, both priced above.
Versions 24.3 / 24.8 LTS / 25.1. Failover is automatic via the Keeper quorum and best-effort, not zero-data-loss. No PITR on ClickHouse.
Billing is hourly against prepaid credits and pauses at a zero balance.
That is about $6,700 a year back, at the same memory and the same availability.
A CPU-parity build (two r7g.xlarge, 4 vCPU / 32 GB each, doubling memory too) prices at about $599 a month, still 33 percent under the Scale bill.
Read before you copy this
Where this math
bends the other way.
ClickHouse Cloud scales compute to zero and resumes in seconds; storage keeps billing. An internal BI service used a few hours a day can genuinely cost less there. This case is the opposite shape: dashboards that answer customers all day.
Their storage is shared object storage at $25.30/TB; ours is EBS per node, doubled under HA. Around 1.6 to 1.7 TB of compressed data the two totals cross. Under 1 TB always-on, the dedicated instance wins decisively. Past it, run both numbers before moving.
HIPAA and PCI live on ClickHouse Cloud's Enterprise tier and their vertical autoscaling has no equivalent here: you resize deliberately instead. If either is a hard requirement, they are the right vendor.
Want this table for your own bill?
Two clicks get you our verified numbers for your shape. Add your setup and a founder replies personally, usually within about two working days, with the same line-by-line format you just read.
Based on this page: ยท
On now or still deciding?
What's the bill you're looking at?
Roughly how much a month, all in?
Matched to RDS, we bill 3.0x less in ap-south-1 (Mumbai) and 2.1x less in us-east-1.
2 vCPU / 4 GB with 20 GB: $24.54/mo in Mumbai, $35.27/mo in us-east-1, against an estimated $73.75/mo on RDS.
See a real RDS bill, before and after →Neon meters compute hours and sleeps. We bill one fixed hourly rate and never sleep.
{price} Same number in a quiet month and a launch month.
Price your size →Two ways off the Supabase bill and you can keep the Supabase stack.
Self-host Supabase on a project server from about $0.02/hr or move just the database to a dedicated Postgres. {price}
How self-hosted Supabase works here →An enterprise client is moving production ClickHouse to us at up to 60% less.
Worked example: off ClickHouse Cloud Scale to always-on managed ClickHouse, $891.86 to $364.03 a month, 59%.
See the ClickHouse bill, before and after →A PaaS bills per service. A project server is one bill for everything on it.
Worked example: a studio consolidated client hosting onto project servers, $390.00 to $82.66 a month, 78%.
See that bill, before and after →A managed database for as low as $5 a month. Dedicated Postgres or MySQL on AWS, with restore, pooling and Multi-AZ failover when you want it.
{price} Postgres and MySQL price the same.
Price your size →A managed database for as low as $5 a month. Dedicated Postgres or MySQL on AWS, with restore, pooling and Multi-AZ failover when you want it.
{price}
Price your size →A managed database for as low as $5 a month. Dedicated Postgres or MySQL on AWS, with restore, pooling and Multi-AZ failover when you want it.
{price} Pay-as-you-go credits, top up from $5, no plan tiers.
Price your size →Most setups map onto two shapes: dedicated databases on AWS and a project server for everything else.
{price} Project servers from about $0.02/hr.
Price your size →Under $200 a month, the calculator beats a human reply. Price it there first.
What's it for?
Where do your users mostly sit?
Dedicated managed Postgres or MySQL, with point-in-time restore, pooling and Multi-AZ failover when you want it.
{region} If you need scale-to-zero or instant branching, Neon fits better than we do.
What managed Postgres includes →Managed ClickHouse for events at volume or TimescaleDB on managed Postgres for time-series next to your app data.
Worked example: off ClickHouse Cloud Scale, $891.86 to $364.03 a month, 59%. {region}
See the ClickHouse bill, before and after →pgvector on managed Postgres: one database instead of a second vendor.
{region} TimescaleDB is one click away on the same instance if you also have time-series.
When a dedicated vector DB is worth it →A project server from about $0.02/hr: push to GitHub and it deploys, several apps and their databases on one box.
Auto-deploy on push, PR previews, custom domains. The $5 welcome credit covers the start of it, no card and you can pause it to $0.
How Projects work →Project servers: many sites per server, one bill.
Worked example: a studio consolidated client hosting, $390.00 to $82.66 a month, 78%.
See that bill, before and after →Most things fit dedicated databases on AWS plus a project server for the apps.
{region}
See it in the live demo →What are you trying to do?
Which editor?
We migrate you, free, on every plan size.
An engineer runs it with you: match the instance, replicate or dump and restore, test the restore, cut over.
See one migration, step by step →The Supabase stack, one click, on a server of your own.
One click from the template, on a project server from about $0.02/hr. Nothing to build or wire up.
The Supabase deploy guide →Paste one config and your editor can run the whole backend.
Run it once, sign in when the browser opens and all 189 tools register.
An hourly instance you pause when done.
Restore a snapshot into a second instance, point staging at it, pause it when idle.
Price a small instance →Push to GitHub and it's live: auto-deploy, PR previews, custom domains.
A project server from about $0.02/hr runs several apps and their databases.
Deploy guides by framework →The one that matters most:
A managed database for as low as $5 a month. Project servers from about $0.02 an hour.
The smallest dedicated Postgres, a t4g.nano with 1 GB, is about $5.34 a month in ap-south-1 (Mumbai) and about $6.41 in us-east-1 and the $5 welcome credit covers its first days. Pay-as-you-go credits, top up from $5, no tiers.
Price your size →When production needs help, you talk to the engineers who run it and a founder reads every request.
Outages, restores, performance issues and migrations reach the people operating the platform. No response-time promise or uptime percentage we cannot prove.
How support and recovery work →Failover, point-in-time restore and backups cover the common ways a database fails.
Multi-AZ failover is a toggle on every engine, Postgres and MySQL add point-in-time restore and all engines get automated backups, live metrics and alerts.
The reliability and recovery details →We migrate you, free. An engineer plans and runs the move with you.
Match the instance, replicate or dump and restore, test the restore, cut over. Any plan size, no extra cost.
See one migration, step by step →Run it in your own AWS account with BYOC, in the region you need.
Encryption in transit and at rest on every managed instance. We do not hold SOC 2 or HIPAA today and we will say so before you spend an hour.
What we can and cannot certify →Fair enough. The live demo is the fastest way to see it.
Or claim $5 to start, no card needed.
Watch the live demo →Our founder
You've already sent us your setup from this browser. A founder will reply to it; no need to send it twice.
Got it. I'll reply to within two working days, from my own inbox.
Noted. One note when it changes, nothing else.
Methodology
-
ClickHouse Cloud rates are their published us-east-1 list prices, checked August 2026 and quoted per line above. Their region-specific rates elsewhere are not publicly documented; this study prices both sides in us-east-1 so a skeptic can recompute it.
-
Selfhost.dev rates are our all-in billed rates, reproducible in the pricing calculator. Both sides priced at 730 hours, in USD.
-
From ap-south-1 (Mumbai), where Graviton runs about a third cheaper, the entry-shape comparison lands near 3x; the region-stamped pair is on the ClickHouse Cloud comparison page.
Frequently asked questions
Doesn't ClickHouse Cloud scale to zero?
What happens past 1 TB of data?
Is failover zero-data-loss?
Can we run this inside our own AWS account?
In short: an analytics team serving always-on, customer-facing dashboards moved from ClickHouse Cloud's Scale tier (two 16 GiB replicas, $891.86 a month at published us-east-1 list rates) to Selfhost.dev managed ClickHouse (two r7g.large nodes plus a Keeper quorum, $364.03 a month at our all-in billed rates, including the per-node boot volumes and the public IPv4). Same memory, same high availability, same region, 59 percent less. The advantage holds for always-on workloads under about 1 TB; bursty workloads and multi-TB datasets can favor ClickHouse Cloud and the page says so.