Status page

A public status page for your services. Polls a list of HTTP endpoints, shows uptime and latency, and exposes a JSON API. One Go file, no dependencies, no database.

Deploy it on velixir and you have a status page on your own domain in about a minute.

What it does

• Polls each monitor on its own schedule, starting immediately rather than after one interval
• Keeps the last 60 checks per monitor in memory and draws them as a bar strip
• Uptime percentage and last latency per service
• GET /api/status returns the raw check history as JSON
• GET /healthz for the platform health check
• Refreshes itself every 30 seconds, with no JavaScript

Configuring it

Two options. The file is the default; the environment variable wins when it is set.

monitors.json in the repo:

{
  "title": "Service status",
  "monitors": [
    { "name": "Website", "url": "https://example.com", "expectStatus": 200, "intervalSeconds": 60 },
    { "name": "API", "url": "https://example.com/health", "intervalSeconds": 30 }
  ]
}

A MONITORS environment variable holding the same JSON. This is the one to use on velixir: set it on the app's Environment tab and you can repoint the whole page from the dashboard without touching the repo.

Field	Default	Notes
name	the URL	Shown on the page
url	required	The endpoint to poll
method	GET
expectStatus	200	Anything else counts as down
intervalSeconds	60	Floored at 5

Running it locally

go run .

Then open http://localhost:8080.

Deploying

velixir deploy

What it deliberately does not do

History is in memory and resets when the app restarts. That is the trade for zero dependencies and a build measured in seconds. A status page is mostly read for "is it up right now" and "was it up an hour ago", and 60 checks covers that.

If you want history that survives a deploy, bind a managed Postgres on velixir and persist checks: it is a contained change to store.record plus a load on startup, and nothing else in the file has to move. DATABASE_URL is already in the environment once you bind one.

It does not alert. No email, no webhooks, no paging. Adding a webhook call in watch when a check flips from up to down is about ten lines, and is the usual next thing people want.

Scale to zero and monitoring do not mix. An app that sleeps when idle is not polling while it sleeps, so run a status page on a plan that stays warm. On velixir that means any paid tier rather than the scale-to-zero Free plan.

Licence

MIT.