v0.44.0
July 12, 2026
Human work that’s left waiting stops silently disappearing: multi-channel alerting with escalation is born. An assigned task or an unanswered "human" status generates reminders via email, webhook or event, following a configurable policy (a ready-to-use default is already in place), respects the tenant’s quiet hours, and can be acknowledged without logging in via a signed link or straight from the bell. Any flow can also be excluded from statistics and have its own retention. Contains a database migration.
-
Feature
No one is left waiting without you knowing
**Step escalation.** An alert policy is a sequence of reminders — email, webhook (via a reusable connector or a manual HMAC-signed URL) or system event — each firing N minutes after the wait begins. The default policy (already seeded for every tenant): email to the assignee after 30 minutes, again after 4 hours, then email to Admins after 24 hours, and finally an `alert.unanswered` event **that flows can intercept** — so a prolonged wait can trigger automations (reassignment, notifying a manager, opening another flow). **Every flow can have its own policy.** From Settings → Notifications you create custom policies (steps, recipients — assignee, specific user or role —, digest window) and assign them to individual flows from its properties; with no assignment the tenant’s default policy applies. **Anti-fatigue.** Multiple reminders for the same recipient in the same window are grouped into a single digest email, instead of one email per pending activity. **Quiet hours.** The tenant can define a window (days + time) outside of which email and webhooks don’t fire — they reschedule to the next available time; in-app notifications stay instant.
-
Feature
Acknowledged without leaving the app
**Ack from a signed link.** Every email/webhook reminder carries an "acknowledged" link (HMAC-signed, no login required): one click suspends the escalation for that activity. **Ack from the bell.** In-app notifications tied to an alert have a direct button to acknowledge them without leaving the interface. **Resolves itself.** Completing the task or moving the item forward automatically closes the wait, with no need for an explicit ack.
-
Improvement
Reporting and data governance
**Response times per operator.** In Settings → Notifications, a report shows how many alerts were acknowledged/resolved and how long it took, per person. **Exclude from statistics.** A flow (e.g. a service or test one) can be excluded from analytics and dashboards with a dedicated flag; the same pass also fixed the system flows that were polluting the aggregates without anyone asking for it. **Retention per flow.** Every flow can define after how many days completed items get permanently deleted (hard delete), for those who don’t want to accumulate history indefinitely. **Folder filter.** Dashboards and analytics can now be restricted to a folder (and its sub-folders) instead of always looking at the whole tenant.