They're not playing the hapless victim. They're doing exactly what the open source license was designed to do and support. They're forking it because they don't like the direction of the current mainters, and are releasing their own open sourced updates to it.
We've seen this happen repeatedly through many open source projects over the last couple of decades. Take the Jenkins/Hudson split, for example, or OpenWrt/LEDE, Node.js/IO.js, or to stretch even further back Mambo/Joomla.
The whole purpose of the open source license is to ensure that you are always able to patch, support, and maintain what you're running. That you're not stuck depending on a single company to keep your software running. That you're always able to add missing features as you see them, etc. (obviously, the license you choose dictates whether you have to provide those back).
By mixing up proprietary code with open source code Elastic betrayed those central ideals of open source, as have Nginx and any number of companies that have adopted this model. The reason why should be pretty obvious. By building the company model around premium modules, they have disincentivised themselves to fixing or enhancing the open source version of the software. Worse, they're not incentivised in any fashion to even accept patches that open source contributors provide that provide the missing features etc.
By way of example, there's a fix for a common issue in Nginx that has landed in their upstream commercial version of Nginx, that hasn't made it in to the open source version (at least last I saw it hadn't, and the fix had been in upstream for quite a while). Nginx ignores DNS TTLs, and can seriously trip you up unless you happen to know to use a particular combination of variables and statically configured DNS resolvers in nginx to make it DNS TTL for servers it proxies too. The company behind Nginx has had zero incentive to provide that in the open source version, or accept patches that fix that behaviour. That leaves people tripping over the same bug for years on end, until they discover, just like so many before them, those blog posts that detail the behaviour and how to save yourselves from it.
What each of these companies have done essentially makes actions like Amazon's actions here completely inevitable.
While I agree with you, companies should really know what a free-as-in-free-beer license means for an open source product, AWS effectively plays the helpless victim in the linked blog post. If it was a simple "we knew this was going to happen, so we forked it" they would have made a
short announcement instead of a 1500 word essay on the issue.
We've seen this happen repeatedly through many open source projects over the last couple of decades. Take the Jenkins/Hudson split, for example, or OpenWrt/LEDE, Node.js/IO.js, or to stretch even further back Mambo/Joomla.
The whole purpose of the open source license is to ensure that you are always able to patch, support, and maintain what you're running. That you're not stuck depending on a single company to keep your software running. That you're always able to add missing features as you see them, etc. (obviously, the license you choose dictates whether you have to provide those back).
By mixing up proprietary code with open source code Elastic betrayed those central ideals of open source, as have Nginx and any number of companies that have adopted this model. The reason why should be pretty obvious. By building the company model around premium modules, they have disincentivised themselves to fixing or enhancing the open source version of the software. Worse, they're not incentivised in any fashion to even accept patches that open source contributors provide that provide the missing features etc.
By way of example, there's a fix for a common issue in Nginx that has landed in their upstream commercial version of Nginx, that hasn't made it in to the open source version (at least last I saw it hadn't, and the fix had been in upstream for quite a while). Nginx ignores DNS TTLs, and can seriously trip you up unless you happen to know to use a particular combination of variables and statically configured DNS resolvers in nginx to make it DNS TTL for servers it proxies too. The company behind Nginx has had zero incentive to provide that in the open source version, or accept patches that fix that behaviour. That leaves people tripping over the same bug for years on end, until they discover, just like so many before them, those blog posts that detail the behaviour and how to save yourselves from it.
What each of these companies have done essentially makes actions like Amazon's actions here completely inevitable.