Alex Van Boxel 🦋 @alexvanboxel.name

I wasn't a big poster anyway, and mainly followed people in tech... they seem to be moving too 🦋... as will I. I love the fact that my domain name alexvanboxel.name can be my handle.

🗓️ Save the date for #BeamSummit! The community will meet in Sunnyvale, CA from Sep. 4-5 in-person! Soft early deadline of April 30 for those needing VISAs. Submit your proposal: beamsummit.org to share your stories on how and where you are using #ApacheBeam 👀

Great into to WebGPU. The things you can do in a browser.

Create a GenAI-based content discovery platform!🧠🦾 Read our recent blog to learn how to build a real-time scalable content extraction solution using #Beam and other @GoogleCloud GCP components ⤸ tinyurl.com/mw29cddw #GCP #VertexAI #GenAI #textbison #textembeddings

“Go 1.21 is released!” by @elibendersky — go.dev/blog/go1.21 #golang

The @ApacheBeam Go SDK is getting a new runner by default in 2.50.0! I talked about it at this year's Beam Summit and I'm excited to work on making the local Beam experience better, and not just for Go users. youtu.be/0iSwoD5nHpE

Next week is @BeamSummit. Join the community conversation on the "Future of @ApacheBeam"... discuss missing features or what the project needs to succeed. Submit your topic in the form: forms.gle/6t3WZhpnDQ4VSs…

Devoxx Belgium 2023 #TomorrowIsNow 🔥 Call For Papers is open until July 14th 🔥 dvbe23.cfp.dev

You don't want to miss #BeamSummit this year! You will find: 🐝 🟡 Sessions with new use cases from companies using Apache Beam, 🟠 Community-driven talks, 🔵 Technical deep dives, 🟢 In-depth workshops and more. 🎟️ Tickets here: beamsummit.org/tickets/

🚀 Excited about the OpenTelemetry Collector? Learn with @alexvb how to adapt it to work with Apache Beam SDK and process historical data in batch by going through the steps, successes, and failures of making a "Go" program work with Beam. 🐝 beamsummit.org/program/

It's the last few days before the Call for Proposals for #BeamSummit2023 closes. Submit before March 20, 2023!!! sessionize.com/beam-summit/

The #BeamSummit Call for Proposals is open until March 20! Go ahead and submit your talks about your experience with #Beam, #streamprocessing, @ApacheFlink, etc. Questions about what to submit? Feel free to ask here! Link: sessionize.com/beam-summit/

Looking for ways to network across top data teams, showcase your work, and get involved? Submit your CFP for our annual Beam Summit happening this June in New York City! We'd love to hear about your experience with #ApacheBeam. Apply now: sessionize.com/beam-summit

Tomorrow I will be at @fosdem presenting @ApacheBeam for streaming analytics. We will review the classics: watermark, windowing. And the not so classic: stateful functions for non-time based "windowing", ML inference, and cross lang transforms (use Python inference in Java too!)

Assuming you're using @HelmPack in your enterprise for release microservices: What's your deliverable? A versioned...

Because #Beam ❤️ #Kotlin, I created a new #opensource lib called #Midgard to : - Have more concise and expressive code - Remove Beam boilerplate code - Propose more Functional Programming style The link to the #Github repo 👇 github.com/tosun-si/midga… @ApacheBeam @kotlin #data

@gunnarmorling @r39132 It feels like you are trying to do tracing, have you considered github.com/open-telemetry…

@r39132 Yes, something in between. Finer than app level, more coarse-grained than query level. Rather, on the level of web request (which might issue multiple queries), so I know by which request / API method a specific query is triggered.

Question for my #Postgres friends: is it practical to (re-)set the application_name on a per request basis (in a web app)? That way, one could identify the source of a query more easily in pg_stat_activity. Is anyone doing that? Or is it too costly?

#BeamSummit returned for its 6th edition! 🎉 Join us June 13-15 in New York City for our in-person event. We're looking for individuals to join our 50+ speakers from premiere data teams at tech, healthcare, finance, media companies & more. Join us now: ow.ly/S1s950MBmy6

@lostluck @ApacheBeam It's good that you already thought about it. We can discuss it in detail at the next Beam Summit :-)

@alexvb @ApacheBeam If you mean "deploy a job to node x" and everything runs on that node? Very possible. It would need to do Fusion, and have failure tolerance added to be useful there. Once tidied up for upstream, there will be a paired email to the dev list about the design goals and state.

I've now written an @ApacheBeam local runner in Go that can handle Process Continuations (AKA streaming transforms). Now to make the code not terrible and contribute it upstream. It's currently unreviewable.

@lostluck I always found that @ApacheBeam lacked a simple running that could run simple self-contained (single pod) streaming pipelines on a k8s cluster, like --runner K8SRunner. I know that's not your focus, but I wonder if your runner couldn't be the basis of that.

I guess I should also handle all the different windows properly too, rather than lean on the global window. But that shouldn't be too bad with the current scaffolding...