All things software & product, honestly - FrontEnd, BackEnd, DevOps, ML - as long as we're solving massive problems.
My code is used by millions of users globally - Potatoes make me happy
this really depends, you can limit your lock in by using Docker based deployments. But even if you don't, there's always a form of lock in whatever solution you choose to use and there's always a cost to re-engineer the solution regardless of platform. So I believe the lock-in issue isn't that much of a deal, and especially not these days with virtualization and cloud agnostic deployment options available in the worst case.
Cost monitoring,
The tagging system in AWS is extremely important to use consistently such that you can more easily group your deployments and get pretty accurate day to day spend of those deployments. In cost explorer it's then a piece of cake to see which parts of your application incur what costs.
Then budget alerts are there to cover hard number value alerts.
If you're running a multi account setup, use organizations with consolidated billing. All your billing for all sub-accounts goes to one account. Again, the tagging system you've setup will help you sort out exactly what spend comes from where.
Really the tools are there.
Cost monitoring has improved a lot in the past couple of years. From my experience in the past couple of years, AWS spends a significant amount of engineering effort reduce churn due to cost especially for companies that are not fully locked in (😉), the alerting can be set to a smaller window such is 1 hr. Allowing you to monitor cost in really smaller increments. Also you should make the most with the Solutions Architects AWS provide. They bring in a lot of tips on how well you can cut down cost.
If you are using something like Terraform for IaC, there are some plugins that estimate the resource cost before provisioning.
In terms of vendor lock-in it's more or less a design problem and a trade off we make during development/product planning. Has nothing to be done on AWS side. Its a conscious decision we make when opting in.
All things software & product, honestly - FrontEnd, BackEnd, DevOps, ML - as long as we're solving massive problems.
My code is used by millions of users globally - Potatoes make me happy
From humble beginnings at an MSP, I've adventured through life as a sysadmin, into an engineer, and finally landed as a developer focused on fixing problems with automation.
My two concerns about AWS
Cost Monitoring is a really painful problem and trust me there's MANY people who have this issue!
There are some companies trying to fix it, but AWS is so cryptic in its spend sometimes it gets really hard to showcase
Lock-in, also fair, but thats the case with most cloud tools though and why they're valued at such high numbers, the LTV is off the chart
Lock-in,
this really depends, you can limit your lock in by using Docker based deployments. But even if you don't, there's always a form of lock in whatever solution you choose to use and there's always a cost to re-engineer the solution regardless of platform. So I believe the lock-in issue isn't that much of a deal, and especially not these days with virtualization and cloud agnostic deployment options available in the worst case.
Cost monitoring,
The tagging system in AWS is extremely important to use consistently such that you can more easily group your deployments and get pretty accurate day to day spend of those deployments. In cost explorer it's then a piece of cake to see which parts of your application incur what costs.
Then budget alerts are there to cover hard number value alerts.
If you're running a multi account setup, use organizations with consolidated billing. All your billing for all sub-accounts goes to one account. Again, the tagging system you've setup will help you sort out exactly what spend comes from where.
Really the tools are there.
Cost monitoring has improved a lot in the past couple of years. From my experience in the past couple of years, AWS spends a significant amount of engineering effort reduce churn due to cost especially for companies that are not fully locked in (😉), the alerting can be set to a smaller window such is 1 hr. Allowing you to monitor cost in really smaller increments. Also you should make the most with the Solutions Architects AWS provide. They bring in a lot of tips on how well you can cut down cost.
If you are using something like Terraform for IaC, there are some plugins that estimate the resource cost before provisioning.
In terms of vendor lock-in it's more or less a design problem and a trade off we make during development/product planning. Has nothing to be done on AWS side. Its a conscious decision we make when opting in.
Great response Gayan!
Thanks for sharing this! Really appreciate it
Anywhere else that's just called "Sales Engineer"