A status page recipe, born from our own isolation
We set out to self-host a status page for PicoPaaS, running as a PicoPaaS app. It didn't work, and the reason is a good one.
Our own isolation said no
Tenant apps on PicoPaaS are default-deny on in-cluster egress, by design. A tenant workload can't reach the platform's own front door from inside the cluster, which means our status page, being a tenant, couldn't phone home to the control plane. That's the tenant isolation doing exactly its job. A tenant that can't talk to the control plane is a security property, so we weren't going to punch a hole in it for the sake of a status page.
So it became a recipe
Instead of a platform feature, we turned it into a recipe anyone can use. You can deploy a Gatus status and uptime page as a PicoPaaS app to watch your external services: third-party APIs, sites hosted elsewhere, DNS lookups.
There is a ready-made public image, ghcr.io/seaburr/picopaas-example-status. Create an app with it, port 8080, no database or bucket:
apiVersion: picopaas/v1
name: my-status
image: ghcr.io/seaburr/picopaas-example-status:latest
port: 8080The template config checks a website and a DNS lookup. Swap in your own endpoints and rebuild the image, since apps have no config-file mount and the config is baked in. Alert through a webhook (Discord, Slack, ntfy, Telegram) because outbound email is blocked. History is in-memory, so it resets on restart.
What it can't do
A PicoPaaS-hosted status page can't reliably monitor services that are also hosted on the same PicoPaaS cluster, for the same isolation reason it couldn't watch the platform. For those, use an external monitor.
The full walkthrough is in the recipes guide: platform.picopaas.xyz/guide/recipes.