Skip to main content

​
Digest Defaults

Defaults for the digest step in any workflow. Applies to both Dashboard digest and Framework step.digest.

​
Default Configuration

FieldDefault
Typeregular
Look-back window5 minutes
Digest time1 hour
Digest keysubscriberId
A regular digest collects events that arrive within the look-back window after the first trigger, then waits the digest time before sending the aggregated notification. With the defaults, the first trigger waits up to 1 hour, scooping up any matching trigger that lands within 5 minutes of one another.

​
Skip the Digest When

  • severity: HIGH, or
  • critical: true
Critical and very-high-severity workflows must deliver immediately. The digest step is auto-skipped in those cases.

​
Digest Key Composition

The digest key controls what counts as the same digest. Default is subscriberId so each user gets their own digest. Add more parts to the key for finer grouping:
PatternUse case
subscriberIdOne digest per recipient (default)
subscriberId + threadIdConversational flows — one digest per thread/post (comments, replies, mentions)
subscriberId + projectIdPer-project activity feeds
subscriberId + organizationIdMulti-tenant per-org digests
Without +threadId, a comment on Post A and a comment on Post B end up in the same digest. The user sees “5 new comments” instead of “2 new comments on Post A, 3 on Post B”.

​
Conversational Digest Example

For a “comment on your post” workflow: 
Trigger (event: payload.threadId: "post_123")
  ↓
Digest: type "regular", look-back 5min, digest time 1h
  Key: subscriberId + threadId
  ↓
In-App
  Redirect: → thread
Each thread gets its own digest. Comments on Post 123 don’t mix with comments on Post 456.

​
When NOT to Add a Digest

  • Single-event flows (order confirmation for one order, password reset for one request).
  • critical: true workflows (digest is bypassed anyway).
  • High-severity alerts (severity > HIGH).
  • User-specified channel-only flows where the user didn’t ask for batching.

​
When TO Add a Digest

  • High-frequency conversational events (comments, mentions, reactions, follows).
  • Activity feeds (“5 new updates in your project”).
  • Lifecycle nudges that may fire many times in a short window.

​
Cron-Based Digests

Use a cron expression instead of look-back + digest time for fixed-schedule digests (e.g. “every weekday at 9 AM”):
  • Dashboard: switch the digest type to cron and provide a cron string.
  • Framework: pass cron: "0 9 * * 1-5" to step.digest instead of unit/amount.
See framework-integration/references/workflow-and-steps.md for the Framework signature.

​
Common Pitfalls

  1. Forgetting +threadId for conversational flows — comments on different posts collapse into one digest.
  2. Adding a digest to a critical workflow — it’s auto-skipped, but it’s a code smell that hides intent.
  3. Two digest steps in one workflow — not supported. Chain workflows with a custom step (Framework) or a second workflow trigger.
  4. Very long look-back windows — digest delivery feels delayed. Keep look-back ≤ digest time.

​
See Also