01 / HTTP monitoring
Start with a public URL.
Choose Website or API, then enter the public HTTP(S) URL Pulse should check. Set the expected HTTP status, an interval available to your account, and a timeout that gives the endpoint time to respond.
Private-network targets are rejected. After saving, the monitor begins checking; find its first result on the dashboard and recent check history on the monitor detail page.
02 / Response validation
Check the answer too.
An HTTP monitor can have one optional response assertion. Choose Text contains to look for an exact, case-sensitive match, or JSON value equals to compare a typed value at a nested dot path such as database.connected.
{
"database": {
"connected": true
}
}Pulse checks the expected HTTP status first. If a response assertion fails, confirmation retries run before the monitor is marked Down.
03 / Notifications
Choose where alerts arrive.
Connect a channel, then explicitly turn its alerts on. A monitor can keep running without an alert channel. Each channel has its own connection steps and alert switch.
Verify your account email address, then enable Email alerts in notification settings.
Telegram
Open the Pulse bot in a private chat, connect that chat, then enable alerts. Telegram notifications use a private chat only.
Slack
During the private pilot, Slack connects through the Pulse operator's workspace. Connections to arbitrary external workspaces are not open.
Discord webhook
In Discord, open Server Settings → Integrations → Create Webhook. Choose the text channel for alerts and copy the webhook URL. You need the Manage Webhooks permission to create one; see Discord's webhook guide and channel permissions guide.
Paste the URL into notification settings, then save with Discord set to On. Pulse sends alerts for confirmed Down states and when a monitor recovers.
04 / Scheduled work
Know when a job stops.
Create a scheduled-job monitor, choose how often it should run, and set a grace period for delays. Save the generated secret when it is shown; it is only shown once.
After the scheduled work completes successfully, send a secure bearer-authenticated POST to the monitor endpoint, not at job start. Before the first ping its state is Pending. A ping marks it Up; if no ping arrives by the expected interval plus grace period, it becomes Down. The next ping marks it recovered.
curl --max-time 10 -X POST \
-H "Authorization: Bearer $PULSE_SECRET" \
"https://YOUR_PULSE_HOST/api/ping/MONITOR_ID" || trueSet PULSE_SECRET in your job environment. Keep the secret out of the URL and source control.
Ready for the first check?