Create one useful monitor, send a real event, prove the broken path, and leave with a record another operator can use.

Goal

Do not start by monitoring everything. Start with one workflow that a customer, founder, finance owner, or operator would notice if it silently stopped.

By the end of this path, you should have:

one monitor with a real owner path
one accepted healthy event
one controlled miss or failed run
one incident or evidence view that explains what happened
one token, request id, and run id you can find again

1. Choose the workflow

Pick the workflow by consequence, not by technical convenience.

Good first choices:

billing or entitlement sync
nightly backup
scheduled report delivery
data freshness check
AI batch output validation
GitHub Actions deploy or release job

Avoid starting with a low-stakes internal cron. Luota should prove value on a workflow where silence is expensive.

2. Create the monitor

Use the onboarding template that matches the workflow shape:

Workflow shape	Template to choose	Why
A job has a clear start and finish	Billing sync run	You can prove duration, failure, deploy SHA, and output.
A job only needs to say it ran	HTTP cron	One accepted heartbeat is enough.
A worker/script already exists	Heartbeat from script	Adds evidence without changing the runtime model.
An output should stay fresh	Freshness check	Luota watches the output promise, not only the process.

3. Send one healthy event

Copy the generated snippet into the real job path or send the wizard's test signal first. The important proof is that the monitor detail page shows the received event and the accepted timestamp.

If an API call returns 202, the event landed. If it returns anything else, treat it as setup failure and keep the error visible until fixed.

4. Run a controlled bad-day drill

Trust starts after the failure path is visible.

Choose one:

skip one heartbeat
close a run with /fail
let freshness expire
make a temporary webhook alert route return 503

The incident should show what failed, when it failed, what context Luota captured, whether the alert route delivered, and where the operator should go next.

5. Finish the record

Before leaving onboarding, add or verify:

owner
alert route
runbook or recovery note
environment
token location
request id or run id from the test event

Done means

You can answer: "What did this workflow promise, what evidence did Luota receive, what happened when the promise broke, and who owns the next action?"

First monitor quickstart