I would like to present a strategic idea to the Board about bringing CentOS Stream to the Operate First community cloud not only to test and prove itself, but to provide a CentOS Stream-based open production cloud for SIGs and the wider open source ecosystem to test and prove on.
I'll use this issue to gather together useful information and to answer any questions.
Is there room on the September agenda for me to bring a 10-15 min discussion?
@alphacc for the latter question, thanks!
Where can one find more info about said cloud and its intended purpose?
10-15 minutes on one agenda topic seems possible if we can keep it within that limit. That will be aided by getting as much information ahead of time as possible, so that we know what it is that we are discussing/agreeing to.
So ... what does it mean to "bring CentOS Stream to the OF cloud?".
Thanks for the questions, I'll put my answers here. The primary project website hasn't been updated yet with clearer messaging. Links are to more detailed content.
First some context, then some links.
This community cloud is a new and growing environment running according to the principles of Operate First.
Operate First is a concept to incorporate operational experience into software development, by extending development to include operating, testing, and proving code in a production environment.
So this cloud is both a place for open source projects to bring code and a reference implementation demonstrating the Operate First principles in action.
Front page has some intent/purpose content:
https://www.operate-first.cloud/
This manifesto is fairly comprehensive of the original thinking behind the Operate First concept:
https://openinfralabs.org/operate-first-manifesto/
Agreed, and I'll start here. 10 - 15 minutes may be a bit greedy, but knowing how agendas can go, I'm being mindful that even with all the background information, there might be further discussion needed around "why" before we get to "what and how."
So the Operate First community maintains bare metal hosts, and has chosen to run Kubernetes and friends as our first environment.
https://github.com/operate-first/
While there is a clear reason to choose Kubernetes to start, the main limitation on the types of cloud architectures that can be offered is people's time, mainly.
So those hosts are running RHEL, and I reckon everyone is generally happy with having a layer of the infra supported by another part of Red Hat.
My idea is that we all add a new community production region to the cloud running CentOS Stream as host, but everything else the same. CentOS developers and operators would have some as-yet-defined role in this new cloud region, as per the methods of operating first.
Concrete example:
Open Data Hub (ODH) has been operating first since the project was born. They had a series of development successes using an Operate First approach, and when it came time to productize into the new Red Hat Open Data Service (RHODS), the software was more than cloud native from a CI/CD perspective. It was truly cloud native in being developed and maintained as a running service by the development team running actual user's workloads. Those same developers incorporated operation lessons back into the code, making RHODS more ready for day-two operations as a result..
Imagine if the ODH team could have all of their workloads also tested and proven in a new CentOS Stream region. Now add in the upstream ODH development team being closely connected to the downstream RHODS SRE team, and what are the possibilities of being able to catch bugs in ODH on CentOS Stream before the RHODS side is doing their work to get it ready for the next RHEL update? And so on.
Because my plan includes the entire open source ecosystem, I also want CentOS Stream in the Operate First community's open cloud because I want CentOS SIGs in the community cloud. And to then be gravity to attract other projects wanting to test and prove their open source code in a CentOS Stream region.
== Other links
Environments and services for users, who have largely been data scientists so far working on AI/ML projects e.g. https://opendatahub.io/:
https://www.operate-first.cloud/users/support/
We can offer 10min+10min for questions for next Wed meeting, let me know and I'll add it to the agenda.
I'm finishing up these slides, but the link is public now: https://docs.google.com/presentation/d/1u9tcCrn-otu9x3pZA9x6hw07Z0xYiznebcKY9nIZ39Y/
When those are done, I'll see about adding the link and agenda item details to the HackMD doc.
@alphacc Yes, please, I appreciate that. I wasn't able to edit the HackMD doc(I think I edited it correctly now), here is my intention for those 20 minutes; I hope to cover the first 3 items in fewer than 10 minutes to give more time for discussion. I'm particularly interested in the SIG leader reactions.
At the October board meeting, the board requested that this ticket be closed, and that the requester be asked to open specific action-based tickets either here (if they are actually board decisions) or in the Infra tracker, if they are related to specific Infrastructure requests.
The Board wishes to thank @quaid for his presentation, and informing us about this project/effort.
Metadata Update from @rbowen: - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.