3-4 services and a bunch of databases?

Assuming there aren’t any particular scaling or performance requirements, if I were managing something like that, I would almost certainly not use k8s. Maybe systemd on a big box for the services?

I agree with you and I'm always confused when people talk about process isolation as a primary requirement for a collection of small internal services with negligible load.

In addition the overhead and reporting drawbacks of running multiple isolated databases is vastly higher than any advantaged gained from small isolated deployments.

For personal stuff I simply run systemd services and that does scale quite a lot (as in, you can rely on it for more production services) I believe.

My hero.

If I had 3-4 services and a bunch of databases, I would look at them and ask "why do we need to introduce network calls into our architecture?" and "how come we're using MySQL and Postgres and MongoDB and a Cassandra cluster for a system that gets 200 requests a minute and is maintained by 9 engineers?"

Don't get me wrong, maybe they're good choices, but absent any other facts I'd start asking questions about what makes each service necessary.

AWS Fargate is popular among large companies in my experience.

Some of them try to migrate from it to a unified k8s "platform" (i.e. frequently not pure k8s/EKS/helm but some kind of in-house layer built on top of it). It takes so long that your tenure with the company could end before you see it through.

Using cloud platform as a service options. For example, on Azure you can deploy such system with Azure App Service (with container deployment), or Azure Container Apps (very suitable for microservices). For database, you can use Azure Database for PostgreSQL (flexible), or Azure Cosmos DB for PostgreSQL.

This way, Azure does most of the heavy lifting you would otherwise have to do yourself, even with managed kubernetes.

In my home environment I run a VM with docker for that.

In a commercial environment I’d still use kubernetes. But maybe something like k3s or if we are in a cloud environment something like EKS.

Usually with time other services get added to the stack (elastic, grafana, argocd, …)

If you use AWS, it's probably easier to use ECS that takes away some of the complexity from you.

Maybe at first, but once you start building all of the IaC and tooling to make it useful and safe at scale, you might as well have just run EKS. Plus, then you can get Argo, which is IMO the single best piece of software from an SRE perspective.

Docker compose is another option.