Notifications
Notifications alert your team when a batch ends in a non-success state. They are created automatically — no manual setup beyond enabling them per environment. Each one appears in the dashboard, and is emailed to the people who subscribed to where it happened.
How notifications are created
When a capture or diff batch finishes in a state that requires attention, deploydiff creates a notification for that environment.
Notification creation is per-environment. You can enable or disable capture batch notifications and diff batch notifications independently in an environment's settings. Both are enabled by default.
Capture batch triggers
| Batch outcome | Notification created |
|---|---|
| Done (all OK) | - |
| Warn | ✓ |
| Error | ✓ |
Diff batch triggers
| Batch outcome | Notification created |
|---|---|
| Done — Pass | - |
| Done — Warn | ✓ |
| Done — Fail | ✓ |
| Done — Skipped | ✓ |
| Error | ✓ |
| Cancelled | ✓ |
Status lifecycle
Each notification starts as Open and moves through the following states:
| Status | Meaning |
|---|---|
| Open | Needs attention — no action taken yet |
| Escalated | Flagged as high priority; still an open state requiring action |
| Archived | Resolved or dismissed — hidden from default views |
The escalated state is intentionally still an "open" state. Archiving is the only way to mark a notification as resolved.
Escalating also emails the subscribers, once.
Transitions
Open ──────────────► Escalated
│ │
└──────────────────────┴──► Archived
│
└──► Open or Escalated (reopen)
When you reopen an archived notification, it returns to its previous state: if it was escalated before being archived, it goes back to Escalated — otherwise it goes back to Open.
Handling a notification
I'll handle it tells your team that you are on it. Your email address appears on the notification, in every list, so nobody else starts on the same failure and nobody assumes it is being ignored.
- It is not a status. A notification you are handling stays Open or Escalated, and can be escalated or archived as usual.
- One person at a time. If somebody else has it, the button reads Take over: taking it replaces them. If two people click I'll handle it on the same notification, the first one has it and the second is told so.
- Let go gives it back when you cannot finish. Only the person who has it can let go.
- It puts the notification in your For you list for as long as you have it, even in an environment you do not follow or have muted.
- You are emailed if somebody else escalates it, subscribed or not.
Archiving keeps your address on it, as the answer to who dealt with it. An archived notification can no longer be taken or let go.
Audit trail
The notification's page says who did what, and when:
| Field | Says |
|---|---|
| Created | When the notification was created |
| Updated | When it last changed |
| Escalated | Who escalated it, and when |
| Handled | Who has it in hand, and since when |
| Archived | Who archived it, and when |
Each row shows the current state: taking over replaces the previous person, and letting go removes the row.
Archive date and actor fields are cleared when the notification is reopened. Escalation fields are preserved — if the notification was escalated before being archived, it returns to Escalated status on reopen.
Subscribing
A notification is created for the whole team. What reaches you is your own choice, made with the bell on each project and each environment.
| Bell on a project | You are told about |
|---|---|
| Not subscribed | Nothing |
| Failures only | Failures in any of its environments, and anything escalated |
| All notifications | Failures and warnings, same environments |
An environment follows its project unless you set its own bell, which then wins: Failures only or All notifications to hear more or less about that one environment, Muted to hear nothing about it while staying subscribed to the rest of the project.
A failure is a capture batch that errored, or a diff batch that errored or ended with a Fail verdict. Every other notification is a warning.
Your subscriptions decide two things: what the For you tab of your notifications shows, with its count in the sidebar, and what you are emailed about. That tab also holds the notifications you are handling. They change nothing for your teammates, and every notification stays visible in the organization, project and environment lists whether you subscribed or not.
Emails
You are emailed about a notification when you are subscribed to its environment or project at a level that includes it, and you can still see it. The email says where it happened and what the batch came to, and links to the notification.
- Emails are only sent fresh. One normally leaves within a couple of minutes. If it could not be sent within six hours, it is dropped: by then it would no longer be news.
- Unsubscribing. Every notification email has an unsubscribe link. It turns notification emails off for your account, everywhere, and leaves your subscriptions as they are: the For you tab keeps working. Turn them back on in your Profile, under Notification emails.
- Account and billing emails are separate and always sent.
How many people were emailed
A notification shows a mail icon and a number: how many people it was emailed to.
- A number above zero — that many subscribers were written to.
- Zero — nobody was emailed. Either nobody is subscribed at that level, or those who are have turned notification emails off. If a failure should not go unnoticed, this is the one to look for.
- No icon — the notification has not been processed yet. This normally lasts a couple of minutes at most.
The number is taken when the emails are written. It does not change if someone subscribes afterwards.
When a notification is escalated
Escalating a notification emails everyone subscribed to its environment or project, whatever the level of their bell: someone who asked for failures only is told about an escalated warning. Someone who Muted the environment is left out, as is anyone who turned notification emails off. The person who escalates is not emailed.
The person handling the notification is emailed whatever their bell says, a muted environment included.
This happens once per notification. De-escalating and escalating again, or reopening an archived notification that was escalated, sends nothing more.
A second icon appears next to the first with the number of people the escalation was emailed to. It reads zero when the only person concerned is the one who escalated.
An escalated notification also shows in the For you tab of everyone subscribed, at any level, and stays there once archived. De-escalating a warning takes it back out for those subscribed to failures only.
Viewing notifications
Notifications are accessible at four levels:
- Your notifications — from your dashboard, across all your organizations: For you for what you subscribed to or are handling, All for every notification you can see
- Organization — all notifications across every project and environment
- Project — all notifications for a specific project's environments
- Environment — notifications scoped to a single environment
All views default to showing Open and Escalated notifications. You can include Archived notifications by selecting them in the status filter.
Each notification links to its related batch so you can inspect the full capture or diff result directly from the notification.
Visibility
Notifications are visible to org Owners and Admins, and to org Members who have been added to the relevant project (as Member or Participant). Org Members with no project membership cannot see notifications for that project.
See Roles & Permissions for the full access matrix.