A little help.
A clearer next step.

Open your dashboard and choose Add monitor. Enter a name and the public HTTP or HTTPS address you want to check, then save it. The Free plan includes 3 live monitors with checks from every 5 minutes. A new monitor shows Pending until its first result arrives. Open its details to see the result, or use Force check now when available.

New account registration is not open for this installation. Existing users can sign in with their configured method. The owner's one-time setup key is not a customer invitation code.

No. Monitors labeled Sample contain illustrative history. They do not make network checks or send alerts, and they do not use your allowance of live monitors. Add your own public URL to collect real observations.

Scheduled checks need a connected checking service. Check Settings → Monitoring service for its last heartbeat. Closing the browser does not stop a separately running worker, but checks pause if that worker's computer sleeps or goes offline.

Online means a successful HTTP response in 1,500 ms or less. Slow is a successful response taking longer than 1,500 ms. Offline means the HTTP check failed, timed out, or returned an error. Rate limited means HTTP 429: the server restricted the monitoring request, so Sentinel waits before retrying. A monitoring failure does not always mean the website is unavailable to visitors.

Sentinel requires 2 consecutive unhealthy checks before opening an incident and sending its first alert. A single failure stays in check history. A normal response closes an open incident and triggers one recovery alert. A URL change, pause/resume, or long gap restarts an unconfirmed failure sequence. Alert channels must also be saved and enabled.

A website may treat automated monitoring differently from a visitor. HTTP 401 or 403 can indicate authentication or firewall restrictions; HTTP 429 is a rate limit. Check your site's access rules and use an appropriate public health endpoint. Sentinel respects retry delays, and Force check now cannot bypass them. It checks HTTP responses; it does not run a full browser or inspect page contents.

Pending means no result is recorded yet. Check that the monitor is not paused or beyond your plan allowance, that its next check is due, and that Monitoring service is connected in Settings. A rate limit may delay the next check. If the service is connected but results remain overdue, include the monitor name, last check time, and displayed reason in a support request.

HTTPS checks validate the connection, while a separate certificate inspector refreshes expiry and issuer details. Missing or old certificate details can mean that inspector is disconnected. It does not by itself mean your certificate is invalid. Open the monitor for its latest connection result and any TLS error explanation.

Open Notifications and choose Connect channel. For Discord, provide your webhook URL. For Telegram, provide your bot token and chat ID. For email, provide an email automation webhook and a recipient address. Save the channel, enable it, then choose Send test alert. You can test a saved channel even when it is disabled; testing does not enable it. Wait 60 seconds before repeating a test for the same channel.

Accepted means the provider accepted the request, not that a person received or read it. Check the destination channel or chat, bot permissions, and your email provider's delivery history and spam folder. Sentinel's Delivery history shows provider failures or an unconfirmed result. Sample monitors never send outage alerts.

The current email alert channel requires your own HTTPS automation webhook. It receives JSON with to, subject, and text fields and must send the email through your provider. An email address alone is not enough. Support replies are separate: they appear in Help & support and are not emailed.

You configure your own HTTPS webhook for Clear cache, Restart service, or Maintenance mode in a live monitor's Remote actions tab. Sentinel sends a POST with action, siteId, url, and requestedAt. Your endpoint performs the work. A successful HTTP response only confirms that the endpoint accepted the request; verify the result on your service.

Open Settings → Manage data & account. You can download monitor settings, check history, and incidents as CSV files. Check history is retained for 31 days. Eligible Free customers can review deletion and confirm with their email and DELETE after a recent sign-in. Deletion removes the active workspace; existing backups and downloads are separate. The deployment owner is protected, and payment records require a billing review.

Customers do not need a setup key. If you joined with Google, choose Continue with Google using the same Google account. If you use a Sentinel password, choose Forgot your password to check whether email recovery is available. Google-only users recover access through their Google account. In-app support requires access to your Sentinel account.

Open Billing & plan to see current Pro pricing and continue to the payment provider. Manage payment details, invoices, and cancellation through the billing portal. Review the price and recurring terms before subscribing.

Sign in and open Help & support → Your requests. Select the conversation to read replies and add details. Requests are private to you and Sentinel's support owner. Replies are not emailed, so check back here. Support is handled during the beta with no guaranteed response time; it is not an emergency response service. Your support history is removed if your account is deleted.