Bangalore, we’re back! 🚀
Join us for another exciting ClickHouse Bangalore User Group meetup as we dive into building scalable, real-time analytics in the cloud.
📍 Awfis Mantri Commerce, Bannerghatta Road
🗓️ 14th March 2026, 10:30 AM onwards
Spots are limited —
The Agentic Data Stack is an open-source, composable architecture for agentic analytics. Runs on a laptop, scales to thousands of agents.
Three layers, all open source:
Every database has its own mental model, and ClickHouse is no different. We wrote up 13 common mistakes we see when teams are getting started, and how to avoid them.
ClickHouse now has an official plugin on the Cursor Marketplace 🎉
The plugin packages the ClickHouse Agent Skills with 28 best-practice rules, and the ClickHouse MCP server for querying your database directly from the editor.
clickhou.se/4sbByfJ
Agentic analytics has changed the warehouse contract.
Amy Chen walks through how to set up ClickHouse Cloud for AI agents.
clickhou.se/4kTxCxI
Our top priority building Postgres managed by ClickHouse is delivering an always-available, reliable, and operationally observable managed Postgres offering.
Here's what we've shipped so far:
We benchmarked three ways to migrate a 1TB table from Postgres to Postgres.
Postgres managed by ClickHouse launched in January, and the most common question is: how can I migrate my large, existing Postgres without downtime?
pg_clickhouse is the fastest PostgreSQL extension on ClickBench.
We built pg_clickhouse to let you query ClickHouse directly from Postgres. Rather than shoehorn analytics into Postgres, pg_clickhouse pushes down full query execution to ClickHouse.
clickhou.se/4qM2v8u
Postgres + ClickHouse is becoming the default architecture for applications that need both transactional and analytical workloads. With Postgres managed by ClickHouse, the two run side by side from day one. But migrating existing analytical workloads off Postgres isn't trivial.
clickhou.se/4alI7pT
We're open sourcing pg_stat_ch, a PostgreSQL extension that turns every query execution into a structured event and streams it to ClickHouse.
Kaushik Iska walks through how it works under the hood: clickhou.se/4rQ8VEG
🐘 It's Postgres Week at ClickHouse 🐘 Today through Feb 20th
Say hello to clickhouse.build - the open source, agentic CLI that helps you add ClickHouse to your existing Postgres-backed TypeScript application.
clickhou.se/4pgZSvB
Can you 4x OTel thoughput with Rust? Meet Rotel: pushing 460K spans/sec/core (vs 130k)
How? By diving deep into Rust's threading model, resolving allocator lock contention, and reengineering JSON serialization with RowBinary encoding to stream data efficiently into ClickHouse.
Shout out to Todd Stoffel for this independent analytics benchmark 👏
While we love seeing ClickHouse top the charts, benchmarks always leave us with one thought: could we be even faster?
clickhou.se/4oEQktn
58x faster and 3,300x less memory 🤯 ClickHouse v25.8 makes analytics over complex JSON better than ever.
Pavel explains how the new JSON type serialisation formats work 👇
The ClickHouse JSON type was already fast, but v25.8 made it 58x faster and use 3,300x less memory.
The problem: JSON docs with 10,000+ unique paths killed performance.
The solution → clickhou.se/48iNcPb
200k rows, 10k paths:
⚡ 3.63s → 0.063s
💾 12.53 GiB → 3.89 MiB
🚀 What's New in ClickStack – Sept '25
clickhou.se/4gOrZ1H
This release brings:
• Import/export dashboards
• Custom OTel Collector config
• Custom aggregations (full CH power 💪)
• Time-window chunking + delta() functions
When ProcessOut needed a faster, leaner way to power billions of payment transactions, they turned to ClickHouse Cloud ⚡️
Read how switching from Elasticsearch to ClickHouse cut costs by two-thirds and shrunk latency from minutes to seconds 👇
How do you unify trillions of data points, reduce costs 40%, and eliminate downtime? Ask Changan Ford.
Full story:
What do zombies, Bollywood, and romantasy have in common?
Nataly Merezhuk will show how JSON, vector, and full-text search in ClickHouse can be used to map relationships across unstructured data at scale and answer your most imaginative questions!
Don't miss this talk at PyConUK on Saturday!
Final day of SREDay! Dale McDiarmid takes the stage at 15:00 to introduce ClickStack.
Observability vendors built three separate engines because they didn't have a database that could handle logs, metrics, AND traces at scale.
We do. And we open sourced the whole stack.
One crypto exchange's actual Datadog bill: $65M/year 🤯
The observability industry has built tools that are more expensive than the systems they monitor. It's like paying more for your car's dashboard than the engine.
How did this become normal?
Heading to SREDay today? Don't miss Rory Crispin at 15:00.
Store everything, aggregate nothing.
Hear about our journey to 100 PB and 500 trillion rows.
100 petabytes. 500 trillion rows. That's our observability platform at ClickHouse.
Thurs 18th 15:00: Rory Crispin on why "store everything, aggregate nothing" actually works
Fri 19th 15:00: Dale McDiarmid on ClickStack and treating o11y as a data problem, not a search problem
clickhou.se/48mvES6
Can more cores make your database slower? Turns out, what flies at 32 cores can crawl at 480.
clickhou.se/48rGmGS
Intel performance engineers went deep on ClickHouse running on their newest processors - 240+ cores per socket, 480+ cores in multi-socket systems.
Built analytics on Postgres? You know what happens next: works great initially, then queries slow down as you scale.
clickhou.se/3IpjLQQ
Add ClickHouse speed without a rewrite:
- Keep Postgres as source of truth
- CDC to ClickHouse
- TypeScript APIs with MooseStack
- Auto-generated React clients
AI demands more than what traditional databases can deliver. Faster ingestion, sub-second queries, and cost efficiency at massive scale. It's no wonder the world's leading AI companies use ClickHouse.
If you're in SF, you can spot our billboards around the city and on the 101! 👀
The ClickPipes Postgres connector now supports CDC from TimescaleDB 🎉
You can:
• Rapidly move existing state with parallel snapshotting
• Keep hypertables in sync with low-latency CDC
• Handle schema changes without breaking the pipeline
• Work with compressed tables
You can also do:
./clickhouse
(with no args)
And then you'll have a CLI + the database engine running all in process.
📢 ClickHouse’s New Observability Stack Unveiled 📢
#ClickHouse is doubling down on #observability with the release of #ClickStack
Check out my fireside chat with @mikeshi42.bsky.social, Head of observability
@clickhouse.com, co-founder @hyperdx.bsky.social :
👉 horovits.medium.com/clickstack-c...
📢 August episode is out! 📢
ClickStack: #ClickHouse’s New #Observability Stack Unveiled
@horovits.bsky.social hosted @mikeshi42.bsky.social, Head of observability @clickhouse.com
🎧 podcasters.spotify.com/pod/show/ope...
📺 www.youtube.com/watch?v=vdsE...
recap post:
medium.com/p/73f129a179a3
