Digest Notifications
Colony delivers two scheduled digests by email: a daily digest summarising the previous day’s pipeline activity, and a weekly reliability digest with a per-repo 7-day reliability report. Both share the same delivery channels, email recipients, send hour, and timezone settings.
What the digest contains
Section titled “What the digest contains”Each digest covers four sections:
Throughput
Section titled “Throughput”| Field | Description |
|---|---|
| Completed today | Number of issues that reached done state since midnight UTC. |
| In progress | Total number of issues across every pipeline state — the sum of all per-state counts, including completed, paused, and blocked issues. |
| Blocked | Issues in a blocked state (dependency-blocked, failure-blocked, or needs-clarification) that require operator attention. |
| Field | Description |
|---|---|
| Today | Total LLM token spend since midnight UTC, with a delta versus yesterday’s spend. A positive delta (more than yesterday) is shown in red; negative (less than yesterday) in green. |
| Last 7 days | Rolling 7-day LLM spend total. |
Efficiency
Section titled “Efficiency”| Field | Description |
|---|---|
| Avg cycle time | Mean elapsed time from issue creation to PR merge for issues completed today, formatted as hours and minutes. Shown as N/A when no issues were completed. |
| Success rate | Percentage of resolved issues (merged or paused) that merged successfully. Shown as N/A when no issues were resolved. |
Anomalies and recommendations
Section titled “Anomalies and recommendations”The anomalies section lists up to five stuck issues — issues that have been in the same pipeline state longer than expected — with their repo, state, and elapsed time. If more than five stuck issues exist, the email notes the overflow count. The dead-letter queue depth is shown separately.
Below anomalies, a recommendations list offers suggested operator actions. When the pipeline is healthy and no anomalies are present, a single “Pipeline is healthy — no anomalies detected today.” message appears instead.
Where to configure
Section titled “Where to configure”Digest notification preferences are configured per organization, not per repository.
- Open Settings from the left sidebar.
- Under Organization, click Digest notifications.
- Adjust the settings described below.
- Click Save.
Configurable fields
Section titled “Configurable fields”Enable daily digest
Section titled “Enable daily digest”Toggles delivery on or off for the entire organization. When disabled, no email or webhook is sent regardless of other settings.
Default: off.
Delivery channels
Section titled “Delivery channels”| Channel | Description |
|---|---|
| Sends the digest to the email addresses in the recipients list (see below). Enabled by default. | |
| Outbound webhook | Fires a digest.daily event to all webhook endpoints subscribed to that event type. Disabled by default. See Webhooks to configure webhook endpoints. |
Both channels can be enabled simultaneously. At least one channel should be enabled for delivery to occur.
Delivery schedule
Section titled “Delivery schedule”| Field | Description | Default |
|---|---|---|
| Send hour | Hour of day at which the digest is sent, in 12-hour AM/PM format (stored as a 0–23 integer). Hour 0 is 12 AM; hour 23 is 11 PM. | 8 AM |
| Timezone | IANA timezone used to interpret the send hour. The dropdown lists all timezones supported by the runtime; if the browser supports Intl.supportedValuesOf, the full IANA list is shown. | UTC (or your browser timezone on first save) |
Colony checks every hour whether it is time to deliver. When the current local hour in the configured timezone matches the send hour and the digest has not yet been sent on that local date, delivery fires. A digest is sent at most once per local calendar day.
Quiet mode
Section titled “Quiet mode”Skip when there is nothing to report — when enabled, Colony suppresses delivery on days where no pipeline activity occurred. A “quiet” day is one where all of the following are true: no issues were completed, no issues are blocked, no issues are stuck, the dead-letter queue is empty, and only the healthy-pipeline recommendation is present.
Default: off (digest is always sent when enabled).
Email recipients
Section titled “Email recipients”The recipients list controls which email addresses receive the digest when the email channel is enabled. Addresses can be added one at a time; press Enter or click Add after typing each address.
| Constraint | Value |
|---|---|
| Maximum recipients | 50 |
| Format | Valid email addresses only |
Default fallback: when the recipients list is empty, Colony falls back to all org owners and admins. To use a dedicated distribution list, add that address and remove individual addresses.
Role requirements
Section titled “Role requirements”| Action | Owner | Admin | Member | Viewer |
|---|---|---|---|---|
| View digest preferences | ✓ | ✓ | ✓ | ✓ |
| Edit digest preferences | ✓ | ✓ | — | — |
Members and Viewers see the settings page in read-only mode with a note that an admin can make changes.
Relationship to the digest.daily webhook event
Section titled “Relationship to the digest.daily webhook event”The email digest and the digest.daily webhook event carry the same payload — they are two delivery channels for the same scheduled digest. Enabling Outbound webhook in digest settings fires the digest.daily event to all webhook endpoints subscribed to that event type. The email and webhook channels are independent and can be used together.
| Email digest | digest.daily webhook | |
|---|---|---|
| Configure channel | Settings → Organization → Digest notifications | Settings → Webhooks |
| Configure recipients | Email recipients list | Webhook endpoint URL |
| Payload format | HTML email | JSON body (signed with HMAC-SHA256) |
| Quiet-when-empty support | Yes | Yes (same suppression logic) |
For full details on webhook signing, retry schedule, and delivery history, see Webhooks.
Weekly Reliability Digest
Section titled “Weekly Reliability Digest”The weekly reliability digest delivers a per-repository 7-day reliability report to your team on a chosen day each week. It is configured on the same settings page as the daily digest.
Enabling and delivery day
Section titled “Enabling and delivery day”Inside Settings → Organization → Digest notifications, scroll to the Weekly reliability digest card:
- Enable weekly reliability digest — toggle to turn delivery on or off for the entire organization.
- Delivery day — select the day of the week (Sunday–Saturday) on which the digest is sent. The digest is sent on that day at the same send hour and in the same timezone configured for the daily digest.
Default: off.
Delivery channels and email recipients
Section titled “Delivery channels and email recipients”The weekly digest shares the Delivery channels and Email recipients settings with the daily digest. There are no separate channel or recipient lists for the weekly digest.
| Channel | Behaviour |
|---|---|
| Sends the report to the addresses in the email recipients list. When the list is empty, Colony falls back to all org owners and admins — the same fallback as the daily digest. | |
| Outbound webhook | The Outbound webhook toggle applies to the weekly digest as well as the daily digest. However, the digest.weekly-reliability event type is not currently in the subscribable webhook catalog, so no webhook endpoint can subscribe to it. Email is the working delivery channel for the weekly digest. |
Scheduling semantics
Section titled “Scheduling semantics”Colony checks every hour whether it is time to deliver the weekly digest. Delivery fires when all of the following are true:
- The current local day of the week (in the configured timezone) matches the configured Delivery day.
- The current local hour matches the configured Send hour.
- The digest has not already been sent in the current ISO week (tracked per-org via the ISO week key, e.g.
2026-W27).
A weekly digest is sent at most once per ISO week, regardless of timezone edge cases or cron retries. The ISO week is determined by the nearest-Thursday algorithm (ISO 8601): a week belongs to the year containing its Thursday.
Role requirements
Section titled “Role requirements”| Action | Owner | Admin | Member | Viewer |
|---|---|---|---|---|
| View weekly digest preferences | ✓ | ✓ | ✓ | ✓ |
| Edit weekly digest preferences | ✓ | ✓ | — | — |
Members and Viewers see the settings page in read-only mode. An admin or owner must enable the digest and set the delivery day.
Report contents
Section titled “Report contents”Each weekly reliability digest covers the following metrics, computed over a rolling 7-day window, broken down per repository. Full definitions and calculation methods are in the Weekly Reliability section of the Metrics reference.
| Metric | What it tells you |
|---|---|
| Completed | Count of issues that reached done state in the 7-day window. |
| Recovered | Count of issues that hit a blocked or failure state and still reached completion in the window. |
| Cycle p50 | Median cycle time (seconds) from first pipeline transition to issue completion. |
| Cycle p90 | 90th-percentile cycle time (seconds) — measures tail latency. |
| Dead letters | Count of tasks that exhausted all retry attempts without resolution. |
| Retry budget | Number of tasks retried (tasksRetried) and total retry-attempt-index sum (totalRetriesConsumed) across the window. |
| Fleet degraded time | Total seconds at least one Worker reported a degraded state during the week. |