Skip to content

fix(notifications): Group alerts and enforce service access - #216

Merged
nfebe merged 14 commits into
mainfrom
fix/notification-alert-incidents
Aug 23, 2026
Merged

nfebe merged 14 commits into
mainfrom
fix/notification-alert-incidents

Conversation

@nfebe

@nfebe nfebe commented Aug 22, 2026

Copy link
Copy Markdown
Contributor

Group repeated metric alerts into incidents instead of sending one message for every sample. This also closes permission gaps found while testing peer deployments and new services, and keeps the white email logo visible across mail clients.

@sourceant sourceant Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Review complete. No specific code suggestions were generated. See the overview comment for a summary.

nfebe added 14 commits August 23, 2026 06:54
Metric alerts now share one incident per rule and resolve only after every affected
series recovers. Explicit alert targets remain scoped to their intended recipients.
Fleet readers can retrieve deployment details from peers without receiving permission
to mutate those deployments.
Service readers can no longer change DNS records or firewall policy. Object storage and
notifications now have independent access controls, and storage attachments require
access to the selected deployment.
The branded header retains its dark background in clients that ignore inline CSS,
keeping the white logo visible.
Notification delivery now uses the typed path exclusively, keeping the branch clean
under CI lint rules.
Operators and viewers no longer inherit administration access.

Explicit grants remain available for custom roles.
Operators can manage alerts for assigned deployments. Other deployments,
host rules, and notification credentials remain private.
Removing a peer now deletes its generated service credential while preserving
user-managed keys with the same display name.
Module permissions and resource grants now have shared implementation and test requirements.
Fleet service credentials follow their own deployment policy.
Peer lists remain available when the current server has no deployments.
Deployments can verify non-HTTP services without unsafe HTTP probes.
Existing HTTP health configuration remains compatible.
Only the reserved Fleet identity uses its API key as the deployment grant.
Other service identities keep the existing permission intersection.
@nfebe
nfebe force-pushed the fix/notification-alert-incidents branch from acee174 to d7ea678 Compare August 23, 2026 05:55
@nfebe
nfebe merged commit 257924d into main Aug 23, 2026
6 checks passed
@nfebe
nfebe deleted the fix/notification-alert-incidents branch August 23, 2026 06:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant