Monitoring, alerts and status pages

A deploy is the start. Launchpad then watches the app and tells you when something changes.

Uptime checks

Every app deployed from the console is checked every five minutes: the web app, each web process, and any extra URLs you listed under monitor in launchpad.json.

Each check records whether the app answered, how long it took, and what went wrong if it didn't. The project page shows the last 24 hours, 7 days and 30 days, with a bar per day.

Give your app a health route (/api/health is the default) that checks what matters, such as the database answering, and returns a non-200 status when it doesn't. A route that always returns 200 tells you nothing.

Alerts

Add channels under Alerts:

• Email
• Slack or Discord, through an incoming webhook URL

You're told when an app goes down, when it comes back, and when a deploy fails. Launchpad waits for two failed checks in a row before saying an app is down, so a single blip doesn't wake you.

Webhook URLs are encrypted at rest, and messages can't be made to ping a channel, whatever an error message contains.

Public status page

Any project can have a public status page at /status/<name>, showing each part's current state and recent uptime. It lists names and health only: never URLs, logs or anything from your code. Turn it on and off on the project page.

Errors from Sentry

Connect a Sentry read-only token and each project card shows its open issues from the last 24 hours, with the top issues on the project page. This is read-only: Launchpad never changes anything in Sentry.

Analytics

Launchpad can detect which analytics a deployed app actually serves, and turn on Vercel Analytics and Speed Insights for the project. A Google Analytics id can be set per project and is injected like any other variable.

Costs

Costs in the console estimates what your providers will bill you this month, read every few hours with the accounts you connected for deploys. It comes with Builder and Studio, and with a Launch pass; AI spend below is on every plan. You pay each provider directly; Launchpad never bills for their usage.

Provider	What's counted	Warnings
Vercel	Hobby: nothing, it's never billed. Pro: the charges Vercel reports so far, projected to the month's end	The account is paused for exceeding Hobby's limits; Pro is on course for well over last month's bill
Supabase	Free: nothing. Pro: the $25 plan, plus Micro compute for each project after the first	The free plan's two active projects are both in use
Render	Each running service's plan price, per instance	—
Neon	Free: nothing	A free project's database is past 80% of its 512 MB
OpenAI, Anthropic	Spend so far, at the same rate to the month's end (AI spend)	Your AI budget

Anything a provider doesn't report through its API (Render's workspace plan and databases, Neon's paid usage, Supabase's add-ons and egress) is left out rather than guessed, and each card says what isn't counted. Each warning goes to your alert channels once a month.

AI spend

If your app calls OpenAI or Anthropic, connect an admin key (read-only in practice: Launchpad only ever reads the cost report) and the console shows this month's spend, with a budget alert at 80% and 100%.

From the command line

The CLI checks on demand rather than on a schedule:

launchpad check     # probe every health URL now
launchpad status    # what Launchpad manages for this app

Continuous checks, alerts and status pages come with the console.