---
key: retry-a-failed-alert-delivery
title: Retry a failed alert delivery
audience: staff
section: critical-alerts
routes: ["/critical-alerts/failures"]
permissions: ["critical_alerts.incidents.view"]
features: ["critical_alerts.core"]
since: 0.0.1
updated: 0.0.1
status: published
summary: Find calls, texts, emails and webhooks that failed to deliver, read the provider's error, and send them again.
keywords: [call failed, text not delivered, email bounced, retry, delivery error, dead letter, failed notification, did not go through]
related: [acknowledge-and-resolve-an-incident, add-a-critical-alert-destination, run-escalations-and-monitors-now]
---

In the sidebar, open **Critical Alerts**, then **Failures**. The page address is `/critical-alerts/failures`.

Every attempt Nova makes to reach someone is recorded. This list holds the attempts that failed: a number that did not connect, an email that bounced, a webhook that answered with an error.

## Before you start

- Your admin has to give your role permission to view critical alert incidents. Retrying needs permission to operate incidents.

## Steps

1. Use the **Filters** button to narrow by **Channel** (voice, text, email or webhook) and **Incident status**, so you can focus on failures whose incident is still open.
2. Read the row. It shows the incident title, the channel and target, the error the provider returned, when the attempt was requested and when it failed, and which provider handled it.
3. Fix the cause if there is one: a wrong phone number on the destination, a missing sender line, an outside system that is down.
4. Press **Retry**. Nova confirms "Delivery attempt retried." and sends the same message to the same target again.
5. Press **Incident** to open the incident itself and see its **Recent delivery attempts**.

## What you see

- The search box matches the incident, channel, destination and error text.
- Two badges on each row: the incident's status and the channel.
- When the provider gave no message, the row reads "Delivery failed without a provider error message."

## If something looks different

- The same failure comes back after **Retry**: the target is still wrong. Edit the destination (see [[add-a-critical-alert-destination]]), publish it, then retry.
- The incident is already resolved: retrying still sends. Check **Incident status** before pressing **Retry**.
- Escalation steps that have not run yet are not failures; they are waiting on the timer. See [[run-escalations-and-monitors-now]].

[[route:/critical-alerts/failures|Open Failures]]