Production-grade container orchestration
The container orchestrator the industry standardized on, scheduling, scaling and self-healing for fleets of services.
Kubernetes is the container orchestrator: declarative desired state for deployments, scaling, self-healing and service discovery across fleets of machines, with an ecosystem, Helm, operators, service meshes, that industrialized its extension. It is how large platforms run compute.
Why it matters to an indie builder: mostly as a boundary marker. K8s solves problems of scale and org size, hundreds of services, many teams, that a small product does not have, and adopting it early trades shipping speed for operational machinery that sits idle. The honest reasoning for a small team is specific, a customer-mandated platform, a workload with genuinely elastic multi-service scale, or managed K8s being cheaper than the PaaS premium at your volume.
A realistic small-scale setup when justified: a managed control plane, one pool of nodes, Helm charts from the application repos, HPA on request rate, and nothing exotic, no mesh, no operators beyond the database one.
Trade-offs: complexity is the cost, the learning curve is real, and debugging descends into YAML and events. Compare with the PaaS blocks for the default path, Fly.io and Nomad for simpler orchestration, and serverless for spiky loads.
Who it's for
Platform teams running containerised workloads at scale who need orchestration, scaling and self-healing.
Official links
Keep exploring
1 in catalog
Kubernetes runs on
4 in catalog
Kubernetes works with
1 in catalog
Kubernetes is written in
1 in catalog