Martijn Visser

Thank you 😅

Boooooo for hosting this in the middle of summer PTO. Now I can’t join

Yeah I can see it being useful when you’ve implemented your own operator, but I also see users making assumptions how some of the built-in operators work and try to change/fix things where they don’t know exactly what they’re doing

The main challenge is always that savepoint and checkpoints can also contain data that hasn’t been emitted to the sink yet. You’re basically looking at committed and uncommitted data from internal state. Is that really interesting for users? Or just the committed part?

I read this more as changes to what would be the equivalent of the State Processor API in Flink?

Now you can teach me how to make all the memes 🥰😍

Before I’m flying back home from Mountain View, I finally took a long walk across Stanford Campus. What a lovely place

Day 7 of the Advent of Flink 2024: on $rowtime and SOURCE_WATERMARK() github.com/knaufk/adven...

Cloudflare incident on November 14, 2024, resulting in lost logs On November 14, 2024, Cloudflare experienced a Cloudflare Logs outage, impacting the majority of customers using these products. During the ~3.5 hours that these services were impacted, about 55% of t...

While failures are never fun, reading up on RCCAs like this one from @cloudflare.social is just such a good read blog.cloudflare.com/cloudflare-i...

I would recommend to check out Restate restate.dev

I think I just saw a notification today in the community on someone porting the existing one to the target interfaces…

FLIP-440 with ProcessTableFunction cwiki.apache.org/confluence/p... is probably the #apacheflink FLIP I've been most excited about since the Table API & SQL upgrade FLIP. Having the ability to build your own processing function with the benefits of the SQL ecosystem (query planner, optimizer) ❤️

Of course. BlacKkklansman was amazing

TIL that “flink” means clever in Norwegian, brave in Dutch, and nimble in German.

I guess I should be searching for “Apache Flink” when looking for folks to add to the data streaming starter pack.

Evergreen skeet: data definitions are hard, y’all.

#dataBS

👋

The number of great movies that are starring Denzel Washington is just insane. Training Day, The Tragedy of Macbeth, American Gangster, Malcolm X, Philadelphia and Remember The Titans. What an actor

There are days where I would like to have more working hours in a single day, when you’re on a roll. Today is not such a day 😂

We will miss you ❤️

If they push it into Calcite, it might

Product Management at Confluent, Flink PMC/committer. Working on streaming data, Flink/Kafka/Iceberg etc

There are thousands of them here in the fall, it’s a super common thing as a Dutch person but so pretty when you actually take a moment to appreciate them

Mushroom in one of the Dutch forests

Our dog enjoying the outdoors and checking out a mushroom

Happy dog in the middle of the forest

Spent a week in one of the Netherlands most pretty national parks, The Veluwe.

It’s always a trade-off. From my experience, the larger the organization is, the more prefer they prefer to have it under a neutral body so they never have to worry about it.

There are a couple of people working adding it to open source. Same for JSON_SR. Did you consider contributing it back?

I guess that’s why everyone is now on the VARIANT data type hype 😏

You might have superior technology, but I really believe you can’t succeed unless the projects lives in one of the open source non profits. It’s just too risky to have a sudden license change by the owning entity.

Post to the Dev mailing list flink.apache.org/what-is-flin... - Happy to take a look at that ticket when I’m back from vacation and ping people where necessary

As with all OSS, it depends on someone wanting to open the discussion and making the contribution :)

Yeah I always prefer to not fork where possible. I don’t think there is a clear “winner”: Slate is promising, but also not proven yet. The work in the Flink community on 2.0 started before Slate existed too. I hope someone will want to take a stab on running Flink on Slate, and see what happens

Flink has always used a forked version of Flink though