Support Node.js 22

Apps are moving to this, we should be running tests against it.

Bump n-health from 13.1.0 to 13.1.1

Bumps [n-health](https://github.com/Financial-Times/n-health) from 13.1.0 to 13.1.1.
- [Release notes](https://github.com/Financial-Times/n-health/releases)
- [Commits](Financial-Times/n-health@v13.1.0...v13.1.1)

---
updated-dependencies:
- dependency-name: n-health
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>

Migrate to @import statements

These are available in later versions of TypeScript and they're a lot
neater/easier to understand. Let's use them.

Remove all default healthchecks

This is a change that just kept on giving!

So I started out wanting to remove the health check that verifies
whether next-metrics is configured properly. We don't need this any
more, I couldn't care less if an app stops sending metrics 🙈 we will
rethink this in future, the right way is probably a heartbeat metric in
OpenTelemetry.

Anyway, once I removed the metrics health check there were no more
default checks. This means there are no more ticking checks at all as
far as I can tell. This allowed me to remove a bunch of redundant code
from the internals.

I tested this by symlinking it into several different n-express apps and
everything starts up fine, including for apps that define their own
healthchecks.

Remove the Heroku log drain check

We're moving this check. We currently have Problem Hub problems based on
the Biz Ops data for Heroku apps, e.g.
https://github.com/Financial-Times/problem-hub/blob/main/problemDefinitions/HerokuApps/herokuAppCustomerProductsMissingLogDrain.yaml

We may also add log drain health checks to cp-healthchecks if we
decide it's an issue. Generally though apps don't change that frequently
and problem hub is probably fine for this.

Remove the Service Metrics alert

This removes the annoying alert when a service is not recognised. This
causes a lot of churn and there's no need for it. next-metrics now
guesses the service name based on the request URL's host name
(Financial-Times/next-metrics#583) and so this
alert will never fire.

This also allows us to do a lot of simplification because we don't have
to handle undefined services.

For all the removed things (e.g. types/options):
a) these was never part of the public API for n-express
b) we don't seem to be using them anywhere

See-also: https://financialtimes.atlassian.net/browse/CPREL-1123

Merge pull request #834 from Financial-Times/ci-2326-bump-next-metrics

chore: ci-2326 bump next-metrics minor version

Merge pull request #831 from Financial-Times/allow-node-22

feat: allow Node 22

Merge pull request #830 from Financial-Times/bump-next-metrics

Upgrade next-metrics to version 12.11.0

Merge pull request #829 from Financial-Times/upgrade-next-metrics

Upgrade next-metrics ^12.9.0 -> ^12.10.0