Luis P. Galeas

Working on - ship enterprise applications in weeks, not years.

Distributed Systems Engineer.

Previously @awscloud, @lendableuk.

[Chris Bakke @ChrisJBakke]

Starting a new company:

You pay us $10k/mo and we send your competitors an endless stream of invites to be interviewed on 3 hour-long tech podcasts, speaking engagements, founder dinners, and all-day tech events so they can never get real work done and they fail.

[Luis P. Galeas @galeaspablo]

@coltmcnealy Indeed. And when your data plane has 3 way replication, you have to divide your capacity estimates by 3 😭

[Luis P. Galeas @galeaspablo]

2022 euros… the last goal of the tournament, from Chloe Kelly, wins it all.

This euros… the same!! ⚽️🦁🙌

I hope to watch it all unfold identically for a third time in 2029.

[BBC Sport @BBCSport]

Chloe Kelly 🥶

Ice in her veins!

#WEuro2025 #BBCFootball

[REY @AOTY2024]

Only 0.64% of her shows were cancelled out of 621 scheduled concerts across her 6 headlined tours

3 of which was due to a foiled terrorist plot
1 of which was due to a military coup

Completely unmatched work ethic

[Luis P. Galeas @galeaspablo]

@JustDeezGuy It depends on what’s being deployed. But I suppose a red flag isn’t necessarily an absolute “that sucks” so I’m sure you know it’s not always bad.

[Luis P. Galeas @galeaspablo]

@HKilset @GergelyOrosz Great comment. We’re playing with this at work.

But not sure if this will be the breakthrough solution. If the search results necessary for my next feature are a meaningful fraction of my codebase and I must reason across all those results, Im screwed!

[Luis P. Galeas @galeaspablo]

HOT TAKE: LLMs will kill SQL.

No, vibe coding will not get rid of databases.
No, I'm not making a point about querying databases with natural language.

But as programming interfaces approach natural language, data storage will becomes more graph-like. Bye SQL!

[Luis P. Galeas @galeaspablo]

@MarcJBrooker @akshatvig This is why I always liked @daniel_abadi’s PACELC framing. It points us back to what’s happening most of the time, ie most of the time there are no network partitions.

[Luis P. Galeas @galeaspablo]

@nathanmarz Yup. And must be the fully diluted number.

[Luis P. Galeas @galeaspablo]

@ArkadiuszKondas yes

[Luis P. Galeas @galeaspablo]

@sandro_vol @ThePrimeagen And the ridiculous stitch pattern! Where the number of interdependencies between services becomes n^2, where n is the number of services. Just no!

[Luis P. Galeas @galeaspablo]

@vietdle you mean 4.5

[Luis P. Galeas @galeaspablo]

Ambar's data streaming abstractions are best in class. And they're only possible because we have some of the best engineering talent in the world. E.g., @Marcelolaza

[Marcelo Lazaroni @Marcelolaza]

Today was time to flip the prod switch on the Haskell projector at https://t.co/jt9Nl5njr0.

It’s now critical infra, processing in real time a few hundred million GBP worth of loans per month among other stuff.

[Luis P. Galeas @galeaspablo]

@xinyubest @jarkwu Does this result in a similar tradeoff to Warpstream when using Express? Ie pay for the cost of frequent puts or deal with increased latencies.

(Assuming I want safety, ie fsync)

[Luis P. Galeas @galeaspablo]

Me: “…stochastic parrots”
Family: “what are sarcastic carrots?”

😂🤣

[Luis P. Galeas @galeaspablo]

@ahachete if quorum:
ok

[Luis P. Galeas @galeaspablo]

@gwenshap So half coffee, half null pointer exceptions? 😀

[Stanislav Kozlovski @kozlovski]

I am incredibly excited to announce that after years of contemplation, I am finally launching an AI startup that will revolutionize the way we do ...

just kidding.

I am launching a newsletter around Apache Kafka & event streaming.

1. What
2. Why
3. How

👇

[Stanislav Kozlovski @kozlovski]

- 3 AM page 🚨
- Kafka broker is running out of disk (11.8TB out of 12TB used) 🚨
- You check everything: throughput, connections network. Nothing seems out of the ordinary 😥
- Retention settings are set to 1 day. It's one topic that hasn't been cleaned in weeks 🤷‍♂️
- You frantically Google, page, and ping team-mates 🙏

... until you figure out ...

the `log.message.timestamp.type` setting

This setting defines what the timestamp on a Kafka record is for storage purposes on the broker.

This timestamp is used, among other things, in the calculation of log retention.

That is, when you have `log.retention.hours=24`, Kafka will delete a log file only when the LARGEST (latest) timestamp of any record in that log file is older than 24 hours.

Back to the `timestamp.type`. There are two supported values:
🔸 LogAppendTime - the time at which the record was appended to the log (disk)
🔸CreateTime - the time that was provided by the Producer when constructing its message

The DEFAULT is CreateTime.

This means that if you have just one Producer client produce a message with a timestamp in the future -- e.g "01-01-2120" - your log file will never get deleted!

This is a common gotcha and can cost you a night's (or a few) sleep.

Ways to circumvent this:
💡- changing `log.message.timestamp.type` to AppendTime
💡- setting size-based retention limits too - `log.retention.bytes`

This will ensure you have a maximum upper bound per topic that you'll never breach 👌