Closed oicheryl closed 1 year ago
Cheryl, I've done this type of project before. Here are a few suggestions:
And, in response to a tweet from @caniszczyk, I think the results should be made publicly available with the CNCF providing value to its members by only letting its end user community participate in making the ratings. Of course, I don't have insight into the membership sales pitch, so take that idea for what it's worth.
@LawrenceHecht
This is an activity meant for CNCF End Users who are actually running these projects in production and share this information amongst each other in a safe environment where there are no vendors, it's by design we do it this way. It is a membership benefit of becoming an official CNCF End User :)
@LawrenceHecht I appreciate the comments!
1.Make sure to have a link to the actual tool/product being evaluated. Make sure to time stamp the assessments.
+1, it's definitely a point in time exercise.
Review the quarterly assessments at a meeting of CNCF end users. This way, you are not assigning a rating only based on the number of votes a technology initially gets.
The CNCF end users review the draft before it is released. Each radar will feature a different set of projects depending on the use case, so each technology will be assessed anew each time.
Don't compromise on making this an invite-only activity.
Also +1. I've already had people ask how they can get their project on the radar.
I think the results should be made publicly available with the CNCF providing value to its members by only letting its end user community participate in making the ratings.
Some of the end users don't have legal/PR permission to publicly state what they use, hence CNCF can publish only aggregrated and anonymized results.
I like the content, but PDFs are not a useful output medium for this kind of thing (imho).
@travis-sobeck you mean the graphic? What don't you like about it?
@oicheryl The problem with a PDF is that it's not interactive. If I'm looking at a graphic with data points, I want to be able to click on the data points to see the source, think of a Grafana graph (or anything similar). Or at a minimum, a link to a specific line on a spreadsheet/webpage with the data. Which is the problem with a PDF, its just one monolithic thing. I want to send a link to someone else to a specific piece of info, not the whole monolithic thing. Lastly, a pdf of a spreadsheet is again not interactive. Give people read access to the anonymized data, even if its just json/csv or google spreadsheet.
@travis-sobeck Got it, I'd absolutely love to make it more interactive. As the very first tech radar it's intentionally a bare bones MVP, but we can add all sorts of bells and whistles over time.
I really want to make this a useful resource for the wider community, so very happy to get the feedback and improve!
@oicheryl Yeah, that's fair.
I would be interested to know what end users recommend for monitoring solutions :)
@oicheryl very nice and useful initiative
Would be great to see this for:
Great insight and very useful.
Voting for future ideas:
Love the concept! Flux was buried in my long list of things to check out and this moved it up towards the top. I'd like to see radars for:
@povilasv @gadinaor @KellyGriffin @jcwinchell I've noted your votes - please keep them coming!
+1 for:
Nice job Cheryl!
+1 for:
Great concept!
+1 for :
+1 for:
Hi @oicheryl - this is brilliant!
As discussed in the TOC call (and wanted to capture here), I think it would be very valuable to also have a radar that focused on technology types (as opposed to specific projects) e.g. different runtimes vs serverless or perhaps different types storage (e.g. object vs file vs block vs KV vs database), as that would provide an indicator for where SIGs need to focus on. (This would be similar to the Techniques quadrant in the example on slide 6 in your deck).
In terms of votes for the next focus areas:
Cheers @chira001 and yes, the Radar format could absolutely be extended to techniques or technology types.
It's getting hard to track so let's do this a bit differently. I'll post one topic per comment, and then people can 👍 the ones they want to see. If you think of something that's not listed, you can add another comment.
Incident management
Log management
Monitoring
Performance monitoring
Security
Service Mesh
Serverless
Virtualization
Requirements management
Storage
Regarding the storage one, Kiran Mova the community leader and chief architect of OpenEBS will be speaking objectively about currently storage landscape for specifically cloud or kubernetes native storage in an upcoming webinar: https://www.cncf.io/webinars/kubernetes-for-storage-an-overview/ This could be some raw material. Anyway +1 for storage and per Alex's point above that itself covers different use cases like DB on Kubernetes tend to be on block whereas back-ups tend to feed into object and so on.
Chaos Engineering
The CNCF end users review the draft before it is released. Each radar will feature a different set of projects depending on the use case, so each technology will be assessed anew each time.
As you mentioned: a single radar is a point-in-time exercise but one of the the key benefits comes from seeing the trajectory of those blips over time to get an idea of a project's uptake and stability.
Just my 2 cents but I'd suggest it would be worth standardising on a number of radar topics and revisiting them regularly.
Service Mesh +1
Service Mesh +1
Thanks @epowell101 for sharing, note that the data for this radar comes entirely from the CNCF End User Community. (See "About the methodology" on https://www.cncf.io/blog/2020/06/12/introducing-the-cncf-technology-radar/.) You could encourage OpenEBS end users to join https://www.cncf.io/people/end-user-community/ to make sure your project is represented.
Thanks @rootsongjc @GladiusK - please also 👍 this comment so I can track: https://github.com/cncf/enduser-public/issues/35#issuecomment-651154303
@oicheryl Hi! First of all, I want to say thanks for your article and podcast at "The New Stack Makers". Is there any opportunity to add comments for technology? I found the only presentation with slides, but there are no comments about technologies and I am curious why somebody chooses GitlabCi, but not TeamCity\Jenkisn or whatever. Same thing we can saw in Through Provoke technology radar. After each tool, there is a small article about it.
Hi @Asgoret, this is an example of the comments/small article you mean right? https://www.thoughtworks.com/radar/platforms?blipid=202005041
The Radar is the aggregated opinion of the whole community, so I thought it wouldn't be right to present a single justification when the companies represented can have very different reasons. Some of the companies in the End User Community (EUC) have shared their reasons for choosing one technology over another, but as it is not a legal/PR approved public statement, this information is confidential within the EUC.
If you are at an End User company (don't sell cloud native products or services) then you can join the EUC to get access to this information: https://www.cncf.io/endusersupporter
Kubernetes Installers to use would be a good topic
+1 for storage
Ingress Controller
Cost Allocation
Zalando open sourced their version of the thoughtworks tech radar (which is also open source): https://github.com/zalando/tech-radar
=> You could use this to have a more interactive version of the radar.
@mischapedia Zalando's tech radar was one of the original inspirations! We are building a new interactive portal which will include more information about the projects and the data behind the radar. Stay tuned...
Autoscaling
If you are interested in learning what end users recommend for a cloud native use case, add a comment below or +1 to vote. Examples could be categories from the CNCF Landscape, or industry verticals such as financial services.
Topics will be selected quarterly by the editorial team as the basis for a CNCF Technology Radar.
More information at https://github.com/cncf/enduser-public/blob/master/CNCFTechnologyRadar.pdf