Examining which Prometheus components consume most of the memory, with a specific focus on storage. Going through TSDB performance, processing data, working with label sets, cardinality, and churn. Practical recommendations to optimize the monitoring setup.
Can't tell for all their projects, but Flux seems to be safe. Stefan Prodan joined ControlPlane last month already and keeps developing Flux CD. Its latest release (v2.2.3) was delivered while working in the new company. He is also working on Enterprise Distribution for Flux CD (https://github.com/controlplaneio-fluxcd/distribution) there now.
I guess it was UI, indeed. Whatever "true engineers" think or say, it is an attractive option for many potential users (think not of Ops only, but of Devs, too). Therefore, it brings a bigger audience, i.e. more adopters, more contributors, more integrations, more whatever…
First of all, that's kinda apples and oranges. If kops is basically an installer, Deckhouse is much more a Kubernetes distribution. It's not just about deploying new clusters but about keeping them running all the way.
Kops installs a "bare" cluster; Deckhouse gives you a set of integrated modules for monitoring, authentication, etc. — as much "out of the box" as you might need to run real applications in K8s. Another outcome is that when you upgrade the Kubernetes version in Deckhouse, it also makes all the relevant work with components and modules in use (they are also updated with compatibility issues in mind, thus everything "just works"). That's why it's marketed as NoOps: it aspires to automate all these routine operations [so you can focus on other exciting challenges ;)].
Deckhouse can also be installed on bare-metal servers, provides SLAs & support for those who need.
