ClickHouse Pricing Teardown
In this post, I dissect pricing of ClickHouse (and it's a complicated one).
This post is a part of series on commercial open source software pricing. See full list of articles here.
What is Clickhouse
Clickhouse is a column SQL database designed for very high querying performance. It was built by Russian IT giant Yandex for their Yandex.Metrika web analytics product, but over time they open sourced the code and spun it off as a standalone company. Clickhouse went on to raise $300M and has $2B valuation.
My banking friends always described it to me as "KDB for the poor", but CH is actually so fast it is used as data storage by PostHog. Their lack of connection to Tableau sucks, but aside from that I've found it a great tool for analyzing lots of data.
31500 stars on Github
Pricing structure
Clickhouse can be self-hosted for free under Apache 2.0 license or use as managed solution. Their managed offering incurs a ton of variable costs – both compute *and* storage. So, they offer a very complicated pricing structure.
It has 3 plans with separate prices for compute and storage which also depend on cloud provider and region. AWS offers 8 regions, Google Cloud Platform – 3 regions, and Azure is not supported yet but will soon.
Developmentplans specify a range of monthly prices per instance and compute/storage prices. It is capped at 1 TB storage/16GiB memory, has daily backups with 1-day retention and burstable CPU and 1-day support SLA. Storage is always paid 100% upfront, after that you pay per compute hours. E.g.aws-us-east-2:storage is $35.63/1 TB
compute is $0.216/hour
monthly price varies between $50 (10% load) and $193 (100% load)
Productionplans increase memory limit to 24 GiB and go pure usage-based with unlimited storage. There is also 2-day data retention, 1-hour support, AWS Private Link or GCP Private Service Connect and automatic scaling. E.g.aws-us-east-2:storage is $47.10/1 TB
compute is $0.6888/hour
Dedicatedplans provide unlimited compute and storage along with custom uptime and support SLAs and advanced security. Pricing is hidden behind sales calls.
Does it make sense to pay?
I don't think so. It's hard to imagine that someone has enough data to need ClickHouse and not have someone on the team who can run a Docker.
This post is a part of series on commercial open source software pricing. See full list of articles here.