- Was originally open source Apache license
- Switched to non-open source Elastic license in Jan 2021 [^1]
- Switched to open source AGPL license in Aug 2024 [^2]
Not to defend the license change(s).
[1] https://www.elastic.co/blog/licensing-change
[2] https://www.elastic.co/blog/elasticsearch-is-open-source-aga...
The idea behind free software was the software was free, but you'd sell the support -- installation help (floppies even)
Elastic were on board with that, and it worked
Until a larger incumbent decided to do their own support (fine), but then sell their service.
Not even a price thing. Far easier for me to spend $50k a year with aws than $5k a year with elastic because we already have a relationship and framework with aws.
They've switched to AGPL, which is a great license, they were just too late, they're the poster child of why AGPL is so important (whether it's another or not is another question)
It was too late to stop it.
Then, Elastic whined about Amazon using Elastic under the open source license they used to build their product. They whined that Amazon wasn't contributing enough. So they switched the license to their product. So Amazon took over maintaining the open source software. Doing exactly what Elastic asked them to do.
Sorry, but everything about Elastic, and especially this most recent announcement of layoffs, scream "bad leadership".
AWS was super greedy and honestly I’m glad elastic even survived their aggressive tactics.
In the early days AWS elastic offering was very weak. It had lots of foot guns and operational problems. We tried hard to use an more native elastic offering and would have preferred it, but it didn't exist.
The fact AWS crappy hosted ES gained any market share is more a testament to how bad ES sales practices were than anything else.
ES really missed the mark by not having a simple, self-serve sales model and instead going all in convoluted "contact a sales rep" enterprise model no one wanted to deal with.
Dozen of hours of useless phone calls, and just nonsensical numbers coming out weeks or months later. Only to go back to the well and try again with a new team of reps after the first team inevitably churned out before any real progress was made.