← All articles

Reliability · 8 min read

The 100% uptime mandate is a design constraint, not a slogan

When a retailer tells you their webshop must never go down, the instinct is to reach for redundancy — two of everything, a failover region, a load balancer. That's necessary. It is nowhere near sufficient. The systems that genuinely hold a 100% availability promise are shaped by that promise from the first architectural decision to the last on-call rotation.

Here is what the mandate really asks of you.

It is an organisational guarantee before it is a technical one

Uptime is lost in change windows, ambiguous ownership and slow decisions far more often than in hardware failure. A real availability target rewrites how the organisation works:

  • Every change is reversible, or it does not ship.
  • Someone is unambiguously accountable for the system at 03:00 — and has the authority to act.
  • The business accepts that "we can't deploy that today" is sometimes the correct answer.

Design for the failure you haven't imagined

Redundancy protects against the failure modes you anticipated. The outages that hurt are the ones you didn't. That is why blast-radius thinking matters more than uptime maths: not "how many nines," but "when this fails, how much fails with it, and how fast can we contain it."

The question is never whether a component will fail. It is what happens to the customer when it does.

Make the system observable enough to trust

You cannot defend availability you cannot see. Logging, monitoring and alerting are not an afterthought bolted on before go-live; they are how the team builds the confidence to deploy on a Friday. If a metric doesn't drive a decision, it's noise — and noise erodes the signal that protects uptime.

Security is part of availability

An incident is downtime by another name. Treating security as a parallel concern to reliability is a category error — a compromised system is an unavailable one. The controls that keep attackers out are the same discipline that keeps the platform up.

None of this is exotic. It is the unglamorous, consistent application of a few principles, held over years. That consistency — not any single piece of technology — is what a 100% uptime mandate actually buys, and what it actually costs.

Got a system that has to stay up?

Let's pressure-test it together.

Start a conversation