Kubernetes
Ask me before you choose Kubernetes.
Kubernetes sounds mature, and it is what the big names use. For most early startups it is the wrong choice, not because it is bad, but because it solves operational problems that do not exist yet, at the cost of complexity, money and specialised knowledge. A short check looks at whether it fits your stage, or whether a simpler setup does the same job without the overhead.
Check this before you commit
What I check first.
- 01
Which operational problem does Kubernetes solve for you that would otherwise remain?
- 02
Do you have in-house (or affordable external) knowledge for 24/7 maintenance?
- 03
How does your product actually scale, uniformly or in spikes?
- 04
What do a simple VPS, managed hosting or PaaS cost for the same workload?
- 05
What do you lose if you have to choose again in a year?
Common patterns
What I see happen most often.
- Kubernetes is chosen for scale that does not yet exist.
- Direct hosting cost is small; the people and operations are expensive.
- Simple alternatives (VPS, Heroku, Cloud Run) never get evaluated.
- "Future-proof" gets translated as "complex today".
What the conversation produces
What you leave with.
- 01
Clarity on whether Kubernetes is needed now or can wait.
- 02
A list of alternative setups with rough cost and benefit.
- 03
A growth path: start simple, migrate when the bottleneck shows up.
Further reading
Articles on this challenge.
- 7 min read
You Probably Do Not Need Kubernetes
Kubernetes solves operational problems most startups do not have yet. What it costs, when it is too early, and which alternatives fit better.
- 6 min read
What I Check Before a Startup Goes Live
Going live is not just a technical moment. The checks I run on support, monitoring, payments, security, backups and rollback before the first customer shows up.
- 7 min read
Before You Spend €20,000 Building It, Read This
A practical test for deciding whether your product deserves a development budget yet.
Ready to talk?
Will Kubernetes actually solve your problem?
Sometimes yes. More often than not, no. Kubernetes is a tool, not a goal.