This message was deleted.

Hi Algam, this is very interesting topic and with Infrastructure as Code it’s possible to transition from PaaS to SaaS service. Tech stack plays vital role and it’s the importance piece of orchestration and maturity of automation in your organization. Here are my two cents on this and happy to talk. In terms of developer experience, it’s doable provided your organization has below : 1. Standardization across all deployments. 2. Proper tagging of resources to a very minute details 3. Proper guard rails and governance with tracking 4. Integration with all necessary tools using API 5. Enablement of automated change , incident and request management. 6. Testing frameworks 7. Security standards etc

I think this article by @Anshul Sao shares a great insight on how to rethink architecture and enable both Dev and Ops team to be "super" https://www.facets.cloud/blog/rethinking-architecture-from-unstructured-diagrams-to-structured-requirements On a personal note, when you say that we want developers to find bugs in internal services we publish, if these are services related to platform, I would highly suggest to think about build vs buy advantage. Why build all those services when you can get a ready made platform running in your cloud, made by experts and even debugged by them (who, btw are not going to skip the complex log, rather they have complex logs for breakfast, lunch and dinner). If you check this group, there are multiple such PaaS available depending upon your team size and usage.

Two things we consider when thinking about buying a service: 1. Cost - we have enormous scale, so things can get expensive. We have quite a few Platform Engineers (Two groups) so for some costly service we prefer open-source or in-house. 2. For some infra needs we can't really trust external products as we can trust our platform groups. We do have external services, it just needs be be a good fit in the terms I just mentioned. Thanks for you input, I will checkout the link you sent.

You are right, in certain aspects, you don't want to get locked down to an external vendor. Although, if you talk about scale, most of the companies will provide you a better ROI than internal developers as the platform building and maintenance costs you in terms of developer hours and building a platform might not be your core competency(not talking about your case specifically, but speaking in general). I remember Humanitec has a page of ROI on their website, which is increases as you scale. Similarly other platforms can give you ROI as high as 500%. Personally, I would prefer my platform engineers to work on building infrastructure strategy, working on security, savings etc and these platforms tend to enable your platform teams to do that. Again, it boils down to cost of the platforms (I am not aware of exact costs, but most of these platforms cost less than a DevOps engineer salary), and provide a significant savings in terms of time and efficiency.