There isn't really much point in higher-frequency probing here
as Prometheus is already hitting those targets for scraping.

Rule that looks at a probe success ignoring failed prober_hosts.

Replace the stock upstream prometheus-bind-exporter dashboard
with our own, which provides a more focused view on the metrics
we mostly care about.

Earlier commit passed tests only because we don't set additional
blackbox probers in tests.

Allow setting custom scrape_intervals for services and blackbox probes

See merge request !277

CORS + SSO is problematic anyway, and the use case for this
has found another solution.

It seems application/javascript is not necessarily the only option...

Rename admin-dashboard service to service-dashboard

See merge request !275

Set some float-specific environment variables in containers

See merge request !271

Fix CORS headers for prometheus endpoint

See merge request !274

Previously we were just not using any modules at all, so the
bug was not noticed.

We would like to expose monitoring API endpoints to an eventual admin
dashboard, but to do so we need to tweak the CORS headers - since
there is no way to set them on the applications themselves, we have to
resort to rewriting the headers in NGINX.

The 'sso_cookie_same_site_mode' config variable is available to
customize the default.

Increase bind tcp-client limit from defaultl

See merge request !273

Usable for tagging in logging / tracing.

This reverts commit 10b117c4.

The run scripts might contain sensitive information in the form of
environment variables or arguments to the containers.

Meant to be used as 'ignorecommand' in fail2ban.

Will be useful for IP filtering and eventual zone-based request classification.

Add the 'schedule_with' service attribute, to allow scheduling a
service along another one.