How are you handling infra/services/networking delivery with 100 silos/teams?
My $job recently made some large scale cloud commitments - the usual big main clouds (3 of them, just because one wasn't enough!), big contracts and plenty of buzzwords.
Working within the cloud practice for about a month, we are faced with plenty of cloud/staffing/knowledge/flexibility/technical issues. Both due to our scale and the hard questions that such a large technical undertaking is forcing us to face.
I'm curious to hear about other folks how they are approaching infra delivery at scale?
- Do you deliver "empty" projects in the cloud providers and let the service team figure out the rest? Do you run this as a professional service and assist teams to design / deliver their services. - Do you have a happy/default path where everything runs in EKS/GKE/Openshift? What if a service team wants a FAT VM running Windows 2000 because some obscure business requirement? - How do you manage the vast technical domain of 500+ different services across cloud providers? - Do you enable all of the APIs of the cloud providers? Or do you restrict and have "safe/default" enabled? - Do you have a cloud placement matrix for the cloud providers? - How do you manage costs when you don't deliver the underlying services? - How do you deal with teams that are just starting their cloud journey / early maturity level? - How do you treat day 2 operations? Managed services / SRE? Push the ownership to the service teams? - Do you have a community of practice for the cloud? How is that effort structured?
I've had some interesting discussions with a couple of friends, but curious to expand my perspective.