Inspect and retry a webhook
Find a webhook that failed, fix the cause, and send it again.
Retry and Replay send a stored webhook again. Use this guide to choose one, recover a webhook after a paused route, an offline CLI or an app error, and see a destination’s backlog.
Prerequisites
Section titled “Prerequisites”- The Docs demo route from the quickstart.
- Your app answering
POST /webhookson port 3000 with a2xxstatus. - The CLI forwarding to your app:
hookspot listen --forward-to http://localhost:3000Choose Retry or Replay
Section titled “Choose Retry or Replay”Three actions send a webhook again:
| Action | Works on | What it does |
|---|---|---|
| Retry on a request | A Rejected request | Checks the request against its source’s current verification settings, then routes the same request again. |
| Replay | Any request Hookspot has routed | Routes a copy of the request as a new request. The original stays as it was. |
| Retry on a delivery | One delivery | Sends that delivery again, on its route’s current path. |
- Retry a Rejected request after you fix the cause, such as a paused source or a missing route.
- Replay a request to send it through your routes again, for example to a new route or after you fix your app.
- Retry a delivery that failed or was skipped to send only that one again.
On Requests, click ⋮ (Request actions) at the end of a request’s row to open its menu. The menu offers Retry on a Rejected request and Replay on any request Hookspot has routed. The request’s side panel and full page show Retry and Replay as buttons beside Copy as cURL.
Each action counts differently toward your monthly deliveries, as Deliveries that count lists.
Retry a rejected request
Section titled “Retry a rejected request”Hookspot rejects a request when no route delivers it. Pause the route to see this, then retry the request.
- Open Routes and click the check mark on the line from Docs demo to
/webhooks. - Click Pause. Docs demo shows an alert icon, whose tooltip reads Requests to this URL are saved but not delivered anywhere.
- Press
tin the CLI to send a test request. Your app gets nothing. - Open Requests and click Needs attention. The new request is Rejected with No route.
- Open Routes, click the paused line and click Resume.
- In Requests, click ⋮ at the end of the rejected request’s row.
- Click Retry.
The request becomes Delivered, and your app gets it. It keeps its ID, so no new request appears.
If you pause the source instead, its requests are Rejected with Source disabled. Click Resume source on the source’s page, then retry them the same way.
Retry first checks the request against the source’s current verification settings. A request Rejected with Verification failed stays Rejected until you correct those settings, such as the secret.
If you retry before you fix the cause, Hookspot rejects the request again, and the rejection counts. Hookspot refuses Retry while your organization is at its hard limit.
A request Rejected with Source deleted cannot be delivered. Hookspot rejects it again on Retry and on Replay.
Replay a request
Section titled “Replay a request”Replay sends a copy of a request through your routes as a new request. In Requests, click ⋮ on the row of the request you retried above, then click Replay.
Hookspot opens the new request, and your app gets the webhook again. The new request shows Replayed from with the original’s ID, as a link to the original. The original and its deliveries stay as they were.
The copy goes through the source’s current routes, so a route you added after the original arrived gets it too. If the source or its routes are paused, Hookspot rejects the copy, as it would a new webhook.
The copy keeps the original’s verification result, and Hookspot does not check it again. So a copy of a request that failed verification still goes to your routes.
Hookspot refuses Replay while your organization is at its hard limit or over its daily data limit. It does not replay a request older than your plan’s retention period.
The CLI’s r key replays a request only to your app, and Hookspot records
nothing. See Replay a request in the CLI
docs.
Retry an offline delivery
Section titled “Retry an offline delivery”You can recover a webhook that no listener received. Its delivery has no attempt until you start the CLI and retry it.
- Press
qto stop the CLI. - Open Overview and click Send test request.
- Open Deliveries. A banner reads 1 delivery skipped: no listener received it.
- Select the new row. It is Skipped with No listener, and its
Attempts count is
0.

No listener received it, so the delivery has no attempt.
- Start the CLI again with
hookspot listen --forward-to http://localhost:3000. The Fix it panel shows Your CLI is connected. - Click Retry delivery.
The delivery becomes Delivered, with one Manual retry attempt. Retry 1 skipped on the banner retries skipped deliveries in bulk, as Retries explains.
Watch a system retry
Section titled “Watch a system retry”Hookspot tries some failed deliveries again on its own. Retries lists which answers it retries and when.
- Stop your app, and keep the CLI running.
- Press
tin the CLI. The CLI showsrefused, and Hookspot records502. - Start your app again.
- Open Deliveries, select the new delivery and click the Open full page icon.
- Select Attempts.

The newest attempt is first.
The Initial attempt is Failed with 502. The next attempt is a
System retry, and it is Delivered with 200.
If your app stays down for every attempt, the delivery stays Failed. Retry it as in the next section.
Retry a delivery
Section titled “Retry a delivery”Retry sends a delivery again, whatever its outcome, as Retries describes.
- Open Deliveries and select a delivery.
- Click Retry.
The Attempts tab gains a Manual retry row. Each retry sends the same webhook again, so your handler can get it more than once.
If the delivery is Skipped, its panel shows Retry delivery instead.
A retry uses the route’s current path. If you fix a wrong path, retry the failed deliveries and they go to the new one.
See a destination’s backlog
Section titled “See a destination’s backlog”A destination’s page shows how many deliveries wait for it and how long the oldest has waited. Send more than your listener can hold to see it.
- Open Routes and click the destination’s card.
- Click Open.
- Send 60 webhooks to a slow handler, as Handle a burst shows.
- Reload the page. The line appears above the delivery counts until your app’s answers make room.

The line appears only while a delivery waits.
For a stalled listener, the line reads like this:
3 queued · oldest waiting 40 min · Docs Demo, dev-laptop| Part | What it means |
|---|---|
3 queued |
The destination’s backlog across all listeners. Above 1,000, it reads 1,000+. |
oldest waiting 40 min |
How long the oldest delivery has waited so far. |
Docs Demo, dev-laptop |
The member and machine of that delivery’s listener. Without a machine, it shows the member alone. Without a listener, the line leaves this part out. |
The age does not forecast when the backlog clears, and a stalled listener’s age keeps growing. With several listeners, the line shows the total and the worst age, so a healthy listener cannot hide a stalled one.
Anyone who can open the project sees the line, viewers included. An open page updates it within about a minute.
To act on it, see Queued deliveries and the backlog email.