> You are going to pay IT staff either way, if it's your own hires (or an MSP), or you are paying via AWS fees. If you're going to pay for IT, why wouldn't you pay for IT that takes orders from you and cares about your business?
Because my IT staff budget would go up if they have all of the additional responsibilities that we currently outsource to Amazon by considerably more than what we would save on our AWS bill by doing it in house. This makes sense because Amazon has a lot more customers besides just me paying for their engineers to manage a given service. So I would be paying a premium "for IT that takes orders from me and cares about my business" and that premium is rarely worthwhile. In the worst case, I have an important customer demo and an Amazon outage occurs--I just explain to my customer that we're affected by the same Amazon outage that's affecting all of their other vendors and half of the rest of the Internet and I forward them the Amazon post-mortem. This has never been a problem for me to date, and I doubt there are many people who have lost deals on account of Amazon in excess of their savings for using Amazon (no doubt there are a few companies for whom this doesn't hold and they should seriously consider on prem).
> Similarly, a single patched Exchange server running on a single VM on a VM host with a UPS backup on it is generally speaking, more reliable and has better uptime than Office 365. Hilariously, Office 365 also costs a lot more.
If I had claimed that every managed service was worth its price for every user, then you'd have successfully refuted me. :) No doubt there are some managed services that aren't worth the price, and even an individual managed AWS service probably doesn't justify the cost of integrating with AWS; however, if you're going to be needing analogs for IAM, Lambda, EKS, SQS, CloudWatch, etc, you're probably going to be spending more on balance (and incurring more downtime) by doing it in house (consider also the difficulty in finding engineers with experience in your exact matrix of tech choices versus general AWS experience).
Because my IT staff budget would go up if they have all of the additional responsibilities that we currently outsource to Amazon by considerably more than what we would save on our AWS bill by doing it in house. This makes sense because Amazon has a lot more customers besides just me paying for their engineers to manage a given service. So I would be paying a premium "for IT that takes orders from me and cares about my business" and that premium is rarely worthwhile. In the worst case, I have an important customer demo and an Amazon outage occurs--I just explain to my customer that we're affected by the same Amazon outage that's affecting all of their other vendors and half of the rest of the Internet and I forward them the Amazon post-mortem. This has never been a problem for me to date, and I doubt there are many people who have lost deals on account of Amazon in excess of their savings for using Amazon (no doubt there are a few companies for whom this doesn't hold and they should seriously consider on prem).
> Similarly, a single patched Exchange server running on a single VM on a VM host with a UPS backup on it is generally speaking, more reliable and has better uptime than Office 365. Hilariously, Office 365 also costs a lot more.
If I had claimed that every managed service was worth its price for every user, then you'd have successfully refuted me. :) No doubt there are some managed services that aren't worth the price, and even an individual managed AWS service probably doesn't justify the cost of integrating with AWS; however, if you're going to be needing analogs for IAM, Lambda, EKS, SQS, CloudWatch, etc, you're probably going to be spending more on balance (and incurring more downtime) by doing it in house (consider also the difficulty in finding engineers with experience in your exact matrix of tech choices versus general AWS experience).