Developer Advocate at QuestDB and all around happy person. Fan of Open Source,Tech Communities,Data&ML.He/him.Ex-AWS,Ex Google Developer Expert. @j@chaos.social
You know it is a major release when you have to create 35 new documentation pages and update another ~75 existing pages. QuestDB 10.0 is out, and it is pretty awesome!
We just shipped #QuestDB 10.0.
QWP: one binary protocol for writes and Arrow reads, 3.6x faster ingestion, 220M rows/sec out.
Live views in beta.
Web Console notebooks, drivable by coding agents.
Parquet tables with schema evolution.
Enterprise 4.0 cold storage ships next.
I'm speaking at Big Data LDN 2026! I will talk about our experience working with agents to build on top of #QuestDB.
If you are in London in September, it is a good opportunity to see what's trending in Data and AI.
invt.io/1txb3uyerft#BigDataLDN
My kid, who glides through netflix, disney+, amazon prime, and youtube, was totally puzzled today when I asked her to change the TV channel. Same device, same remote. Completely different mental schema
These are the type of blog posts I would like to be able to write.
My colleague @AndreyPechkurov goes deep to explain how we (ahem, he) added Window Joins to QuestDB, and why it outperforms other database engines.
WINDOW JOIN in #QuestDB is not just syntactic sugar.
We published a deep dive into how we parallelized and vectorized it, plus benchmarks against Timescale, DuckDB, and ClickHouse.
On the tested workload: 5x faster than the previous QuestDB path.
Link in thread 👇
ALT Dark navy thumbnail with the headline "How we made WINDOW JOIN parallel & vectorized" — WINDOW JOIN in pink, the ampersand in cyan. Below it, a two-stream timeline: a row of bright cyan anchor events on top, and a dense pink data cloud beneath. Eight pink window frames cut vertically from each anchor down through the cloud, with the cloud noticeably denser inside each frame. A small pink pill at the top of every frame labels it T1 through T8, suggesting one thread per window. A line of monospace type along the bottom right reads "8 windows · parallel · SIMD".
We’re excited to be nominated for the DBTA Readers’ Choice Awards.
If #QuestDB has been useful to you, we’d love your support. Voting is open until May 22.
Link in thread 👇
A few weeks ago my colleague @jerrinot noticed an interesting commit in OpenJDK fixing a 7 year old performance issue. He tested it out and found a 400x speed boost. Yesterday the @ThePrimeagen did a rad video featuring his content. Pretty cool! piped.video/watch?v=R3ydGMRt…
How does your database cope with datasets with 6000+ columns? #QuestDB handled them fine, but struggled to copy data across tables. No more. github.com/questdb/questdb/p….
And this is why you *sometimes* need specialised databases
My machine kept freezing when I used async-profile.
This is the deepest I've ever gone: debugging a Linux kernel deadlock, performing live surgery on a frozen CPU with GDB, and pulling off a questionable resurrection.
Read the full autopsy: questdb.com/blog/async-profi…
When I joined #QuestDB in 2022, data ingestion peaked at ~1 million rows (ts+10 strings + 10 longs) per second on a single instance using the TSBS benchmark. 3.5 years later, we are now ingesting 11.36 million rows/sec. Not too bad!
Sure, hardware is a factor. Back in 2022 we ran the benchmark on an AWS EC2 m5.8xlarge instance with 32 vCPUs, and we now use an r8a.8xlarge (32 vCPUs as well). But the real boost is in the revamp of the whole ingestion layer
Acabo de enviar mi propuesta "Desarrollar una base de datos de alto rendimiento en Java es imposible" a @t3chfest, donde hablaré de cómo hacemos #QuestDB. Si eres como yo y lo dejas todo para última hora, ESTA es la última hora para enviar la tuya en t3chfest.es/2026/call-for-ta…
En la agenda del inicio del curso escolar, el alcalde @marecuencopp nos recuerda que si eres de Leganés, estudias aquí, y llegas a la Universidad, puedes llegar a ser futbolista. Perdón por no ilusionarme con esta gran visión para las niñas y niños de Leganés 🧵
VTI, SilcoTech, o ZIV están a la vanguardia de la industria, AEQ, BTESA, o Twincam son referentes mundiales en audiovisual... En las letras tenemos a Santiago Gómez Valverde, Medalla de Plata de la Ciudad, o a Mario de las Sagras (soloeme)...
Pero en el inicio de curso, el señor @marecuencopp ha decidido que para inspirar al alumnado de Leganés a que desarrollen sus estudios en esta ciudad los mejores ejemplos son el fútbol (2 veces) y el Taekwondo. Que están muy bien, pero se me quedan un poco cortos. Y así todo.
On my way to @berlinbuzzwords, where I will be presenting tomorrow about some of the techniques that allowed #QuestDB to perform 6x faster in ~2 years of development.
If you like fast #timeseries databases, come say hi!