Maintenance mode without drama

Let's say your website goes down. Not hypothetically — actually down. You stop the container. What happens to traffic?
With Millau, nothing catches fire. Visitors are instantly served by a fallback service. No timeouts. No reloads. No redeploys.

How it works

In the manifest below, two services handle traffic: homepage and fallback:

services:
homepage:
labels:
- "millau.enabled=true"
- "millau.hosts=company.com"
fallback:
labels:
- "millau.enabled=true"
- "millau.hosts=*"

When the company.com is online, Millau routes requests to the homepage service, thanks to exact host matching.
If company.com goes down, Millau falls back to the next service, which matches any host *.
Bring company.com back, and routing automatically resumes.

Use Cases

But fallback isn't just for outages. You can also:

• show commercial splash screens before launch,

• serve unknown hosts,

• hold users at a soft-launch gate during Friday deploys 😅

If you also don't want an ingress proxy that needs constant babysitting — maybe Millau is worth a look.

millau.net for docs 📖 github.com/codelev/millau for stars ⭐