I can write my own config management, secret management, volume mounting, deployments, replica scaling, hardware affinity, antiaffinity for resilience, rollover kicking out e.g. staging if there's not enough room to schedule production workloads, and the list goes on.
I can also figure out how to install and upgrade each piece of software in my universe, across a few different languages.
Or I can just use kubernetes.
From the perspective of my developers, they tell me a few things like: "I need 4GB RAM, 2 cores, I need access to the postgres secret to talk to it, and here is my hostname" and while they can tell me many, many more things, with that little information, they can have a full application in one of several different languages in minutes.
If I were to hand roll this, it'd be held together by prayers and duct tape. And I lose out on a huge active community building tools around this. I now have istio and get cross service telemetry. For free. I can set networking rules up setting up QOS between arbitrary services, both inside my network and outside. All applied behind a single interface. There's a lot of stuff I get for free, and as my team discovers things it needs to do, I have a consistent layer to do that all behind.
I'm one guy. I don't work for Google. But my team still manages _a lot_ of services, both on application end, and things like queues, databases, etc. You can say we don't need all of that, but you'd be wrong. We arguably could use more. I can reason about my infrastructure in both the small and the large.
Then you get into the idea of transferring to other companies. If someone comes into my org, they can look at what I have and see more or less everything there is in any part of my infrastructure. It's all right there, laid out in YAML. If I were to switch companies, or even teams, it'd be the same. Having a lingua franca of common terms and ideas is super critical there. To say nothing of all the work happening in the ecosystem allowing me to help my users ship features faster to our stakeholders. Doing this in VMs would be a literal nightmare.
So, respectfully, I believe your opinion is not correct here. You personally may not need it, but it's not very long before you get to reasonable amounts of services for any project that's sufficiently complex, in my experience. And even when I have just 1 or 2 I still need many of the primitives of kubernetes.
Exactly. Kubernetes has established patterns. Once you understand them then everything gets EASY. How toñset everything up, how to deploy, networking, service communication, secrets, etc
I've had teams create brand new services in hours that used to take a week. They can create, deploy, debug and manage their own service without any help from me. Their service runs and looks exactly vthe same as every other service. When things go wrong, they know where to look.
My team is smaller than it was last year while running 3x as much. That is all from a good combination of CI jobs, terraform and kubernetes. Also, our cloud bill is WAY down from last year because of the fewer number of VMs we need to run.
Is kubernetes a fad when the CTO is happy because costs are down, while productivity and stability are up?
I think a lot depends on the environment you're working in / migrating from. If you're working from an on-prem datacenter with VM's and you're not moving to the cloud for whatever reason. I think your point of not having to hand roll config managment, secrets management volumes ect.. is a good one and kubernetes makes a lot of sense in that environment. Or using kubernetes as a target for vendor apps that may need to run in unknown or exotic environments.
But if you're already in the cloud most of the k8 solutions already exist as individual tools, with little operational maintenance needed. "I need 4GB RAM, 2 cores, I need access to the postgres secret to talk to it, and here is my hostname" is all achievable with lambda's or ecs fargate task definitions and parameter store. Managed services for queues and databases lower operational burden. All of this while not having to risk a cluster upgrade bringing everything down or now having to increase operational overhead by running multiple clusters to have some operational blast doors.
All of this to say if kubernetes is working for you thats great! Congrats. But I'm on the hill with OP. Most platform/infrastructure engineers that I've heard pitch kubernetes because they are trying to solve one problem I.E secrets management or deployment ect... not all of the problems at once and bring in kubernetes complexity when there is already a fully managed, safer and cheaper solution to migrate to doesn't usually make sense.
There's no 100% right or wrong answer but I think you really have to evaluate what you're migrating from when evaluating moving to kubernetes
There are many options that are not kubernetes and not VMs. Cloud native offerings are extremely simple and powerful. ECS does almost every single thing kubernetes does with less headaches.
To be fair, the OP is probably referring to big companies' stable enterprise applications with a few thousand users. There is a push to migrate everything to Kubernetes but the applications are doing ok on dedicated VMs.
I can write my own config management, secret management, volume mounting, deployments, replica scaling, hardware affinity, antiaffinity for resilience, rollover kicking out e.g. staging if there's not enough room to schedule production workloads, and the list goes on.
I can also figure out how to install and upgrade each piece of software in my universe, across a few different languages.
Or I can just use kubernetes.
From the perspective of my developers, they tell me a few things like: "I need 4GB RAM, 2 cores, I need access to the postgres secret to talk to it, and here is my hostname" and while they can tell me many, many more things, with that little information, they can have a full application in one of several different languages in minutes.
If I were to hand roll this, it'd be held together by prayers and duct tape. And I lose out on a huge active community building tools around this. I now have istio and get cross service telemetry. For free. I can set networking rules up setting up QOS between arbitrary services, both inside my network and outside. All applied behind a single interface. There's a lot of stuff I get for free, and as my team discovers things it needs to do, I have a consistent layer to do that all behind.
I'm one guy. I don't work for Google. But my team still manages _a lot_ of services, both on application end, and things like queues, databases, etc. You can say we don't need all of that, but you'd be wrong. We arguably could use more. I can reason about my infrastructure in both the small and the large.
Then you get into the idea of transferring to other companies. If someone comes into my org, they can look at what I have and see more or less everything there is in any part of my infrastructure. It's all right there, laid out in YAML. If I were to switch companies, or even teams, it'd be the same. Having a lingua franca of common terms and ideas is super critical there. To say nothing of all the work happening in the ecosystem allowing me to help my users ship features faster to our stakeholders. Doing this in VMs would be a literal nightmare.
So, respectfully, I believe your opinion is not correct here. You personally may not need it, but it's not very long before you get to reasonable amounts of services for any project that's sufficiently complex, in my experience. And even when I have just 1 or 2 I still need many of the primitives of kubernetes.