In 2004, the CentOS project started. The idea of the distribution was to be binary compatible with Red Hat Enterprise Linux. This worked well for quite sometime. The Linux community was able to offer up a stable and useful distribution that rivaled corporate operating systems everywhere. One could easily acquire RHEL without the price tag, and many system administrators who had been trained on RHEL could be hired and immediately get to work. Red Hat even benefited from this, as more people were able to test software on RHEL, administrators could practice for their RHCSA and RHCE exams, and with a greater install base Red Hat benefited from the network effect; it was a great thing for everyone. 10 years after CentOS's arrival on the scene, CentOS was acquired by Red Hat. 4 years after that, Red Hat was targeted for acquisition by IBM, and one year later (2019), IBM acquired Red Hat. In 2020, IBM terminated the primary CentOS release.
CentOS 8 will only be supported through 2021. CentOS 7, as of this writing, will maintain support through 2024. CentOS Stream will continue. The general idea is that Red Hat's Fedora will continue to act as a research system and testing ground for RHEL. Updates to RHEL release will then be tested on CentOS Stream prior to their inclusion in RHEL.
Naturally, the timing of this change has not been well received. For this to occur 1 year after Red Hat's acquisition by IBM makes IBM look as if they're embracing, extending, and extinguishing CentOS. It further makes IBM look like they do not seem to know what they're doing: a remake of CentOS is already underway as Rocky Linux.
Ultimately, I am not certain if this is a good thing or a bad thing. On one hand, it is convenient that there is still a gratis version of RHEL. On the other hand it is inconvenient that this version is a testing distribution for RHEL proper. It may not be perfectly compatible with downstream Red Hat releases. Additionally, it doesn't help build/maintain trust with Red Hat that they'd dramatically cut the support timeline for a production operating system release.
For the most part, Linux systems have become, overtime, easier to use and setup but more bug ridden and of poorer software quality. There has been a shift in the way software is developed, delivered, and supported that puts more emphasis on continual release, and effectively devloping in production. As a result, people are getting features more quickly, but they are getting these features in a state that is really only partially functional.
It is sad to see CentOS go this way, but this is symptomatic, in my humble opinion, of a more general shift away from stable software within the industry. All software is now a moving target, and apparently even enterprise operating systems are going this direction.