Hi folks. Is there a platform reference architectu...
# general
l
Hi folks. Is there a platform reference architecture that is held in high esteem ? I thought I would ask some humans first to get some nuance.
l
There are lots. What does your setup look like? This is an example of a reference architecture on AWS using Backstage. The community also has ref archs for multi-cloud, openshift, high security etc
Here is an article I wrote on being able to make your own ref arch diagrams and the thoughts behind the design of these ones https://platformengineering.org/blog/create-your-own-platform-engineering-reference-architectures
t
What are you asking? Do you ‘just’ need the cloud infrastructure, or do you also need the full stack? We are building a reference architecture with Azure, .NET, ReactJS, and GitHub. Just clone the repository, run a single command answering which Azure Subscription and Azure region you want to host in, and in less than 30 minutes you will have a full Azure Cloud Infrastructure following all best security practices with managed identities (no passwords), auto-scaling from 0 to millions of users. It will set up staging & multi-region production environments in GitHub and Azure, and perform passwordless deployments (using OIDC), run tests, and static code review with gated reviews, enabling application deployments in less than 10 minutes. Additionally, you will get a full development setup enabling you to run everything locally with one command with only .NET, Node, and Docker as prerequisites. New engineers will be up and running in minutes. The solution shows how to build a high-quality distributed .NET architecture with a reverse proxy and BFF, with multiple service full-stack solutions. We are constantly evolving the platform, and we do need to document the solution (working on it 🙂).
p
A bunch of good links here that are new to me. Thanks folks!
m
yeah, same here
d
What I'm currently seeing is churn that happens when you try to think of the platform and the portal with too high a degree of separation.
c
Could you explain that a bit more @Dave Dembeck? Do you see teams leaving the platform if you neglect either frontend or backend of the platform too much?
@Lyronne Rangan - have you seen those already? https://humanitec.com/reference-architectures they’re OSS and completely terraformed, so you can get one running in under an hour usually. Also feel free to replace any icons you don’t like in the spirit of the https://platformengineering.org/platform-tooling landscape - they’re reference architectures and meant to be adapted to your requirements. LMK if you have any questions!
c
Hi guys... For the past few months I have been working on a reference implementation of an IDP using best practices. https://github.com/alustan Project is almost up and running though currently resource constraint on cloud testing. You can check out the project. Contributions are wellcome.  If anyone can be of help in sponsoring AWS cloud testing will be greatly appreciated.
Open to work if you are hiring
d
Jay, ( https://platformengin-b0m7058.slack.com/archives/C02DF3KKNES/p1717103439049589?thread_ts=1717048911.360729&cid=C02DF3KKNES ) Problem arises at scale where the platform engineers are isolated from the portal team. These articles are written from the viewpoint assuming those who build the portal understand what the platform actually does.
In a way, this idea that we need to differentiate between a portal and the platform are unhelpful.
c
Hmmm… that’s an interesting thing to happen. Architecturally the portal is part of the “frontend” of your platform, while other components are the “backend”. Any software where frontend- and backend-devs don’t collaborate to achieve a cohesive outcome is doomed.