Notifications

Notifications keep you up to date on relevant events in the workspace without having to check every module manually. You open the overview via the bell icon in the header; similar messages can be grouped so the list stays manageable.

Notification centre

• A click on the bell opens the list of recent messages.
• Unread entries are usually highlighted; read ones appear more subdued, depending on the theme.

Grouping

• Similar notifications can be combined into a single group (e.g. several similar scan hints).
• This avoids long lists with repetitive single entries.

Automatic notification types

The platform can produce, among others, the following automatic notifications:

• Scan completed - the run has finished; you can jump straight to the results.
• Scan failed - the run did not finish successfully; details are in the scan or the message.
• Guardrail warning - a configured protection rule reports a warning event.
• Guardrail block - an action was blocked by a rule; check context and configuration.
• Token warning - hint about low balance or a limit being approached.
• Invitation accepted - a workspace invitation you sent was accepted.
• Referral reward - bonus tokens from the referral programme once the conditions are met.

Mark as read

• Mark all as read - resets the state for visible messages and reduces visual noise.
• Individual messages can also be counted as read depending on the UI when you open them.

Retention and auto-deletion

• Notifications are automatically deleted after 30 days so the history doesn't grow indefinitely.
• If needed, transfer important content to tickets or documentation before it expires.

Day-to-day prioritisation

Which messages first?

1. Guardrail block - can stop critical actions; investigate the cause before continuing.
2. Scan failed - affects data quality and planning; check promptly.
3. Token warning - prevents later start failures with a low balance.
4. Scan completed - lets you know when new results are available.

Quiet times

• If you expect many similar messages (e.g. test runs), use grouping and mark all as read to keep your head clear.

Tips

• For failed scans first read the message, then the scan detail view and check logs or guardrails if needed.
• Take token warnings seriously and top up in time to avoid interrupted runs.
• Remember: after 30 days old notifications are gone - put important facts into tickets instead.