What counts as one interruption?

One contiguous run of awake records inside the main sleep session. Adjacent awake chunks are merged first, so a 20-minute waking recorded as several pieces counts once. Awake time that touches the start or end of the session is latency or the final wake-up, not an interruption.

Why does a user have awake duration but no interruption count?

Their provider reports a total awake time for the night without timestamped awake periods. WHOOP is the usual case. Awake duration can be taken from the total; a count of separate awakenings cannot.

Which one feeds the Sleep Score?

Awake duration, through the sleep_continuity factor. The interruption count is not a factor.

Sleep interruptions and awake time

What the interruption biomarkers measure

Biomarker	What it is	Unit
`sleep_interruptions`	Number of separate awakenings inside the main sleep session	Count
`sleep_awake_duration`	Total minutes awake after sleep onset during the night	Minutes

Both are reported daily over the 6pm to 6pm sleep day, dated by the morning it ends, and both require a wearable.

Field	`sleep_interruptions`, `sleep_awake_duration`
Unit	Count, minutes
Reported	Daily, one value per night, when the source records awake periods
Window	6pm to 6pm, profile local time. Dated by the morning the window ends.
Wearable required	Yes
API	`GET /api/v1/profile/biomarker/{externalId}?categories=sleep` plus one `types` parameter per field and a date range
Webhook	`BiomarkerCreatedIntegrationEvent`. Entries update in place, upsert on `id`.

Data sources

Both need awake data from the night’s winning source, chosen as described under sleep duration.

Source type	Interruptions	Awake duration
Timestamped awake records (Apple Watch, Fitbit, Oura, Garmin, most Health Connect writers)	Yes	Yes, summed from the records
Provider awake totals only (WHOOP)	No	Yes, from the total
In-bed or asleep only, including the phone estimate	No	No

How Sahha calculates sleep interruptions

Interruptions.

Take the main sleep session, the longest in the window.
Collect awake records that fall entirely within it and merge adjacent or overlapping ones into runs.
Discard any run that touches the session’s start or end.
Count the remaining runs.

Awake duration. If the provider reports an awake total for the night, use it. Otherwise sum all awake records in the window.

A source with no awake records produces no interruption count. Absence means the source cannot see awakenings, not that there were none.

Interpreting the values

Devices disagree on what counts as awake. Some flag a few minutes of stillness-breaking movement as awake; others require several minutes. The same night on two devices can differ by a factor of two in count. Compare within a source.

Brief awakenings are normal. Healthy adults spend some time awake after sleep onset every night, and the amount rises with age [1]. A count of five or more, or awake duration above 45 minutes, sustained over a week, is what the continuity factor treats as a problem.

Awake duration includes the wake-up tail on some sources. A user who lies in bed for 30 minutes after waking, still tracked, adds those minutes if the provider counts them as awake within the session.

Score factor. The Sleep and Wellbeing scores carry a factor named sleep_continuity, unit minute, goal 0, whose value is the night’s awake duration. Credit decays with awake minutes:

Awake minutes	Factor score
0	1.00
15	0.93
30	0.87
60	0.75
120	0.56

Users whose source has no awake data have no continuity factor and are not penalised for it.

Trend. A sleep_continuity trend is available on the factor score.

Comparison. None.

Use cases

A fragmentation explanation. When duration is fine and the Sleep Score is not, awake minutes are usually why. “You were awake for 52 minutes across 4 wakings” is the sentence to show.

GET /api/v1/profile/biomarker/{externalId}?categories=sleep&types=sleep_interruptions&types=sleep_awake_duration&startDateTime=2026-08-26&endDateTime=2026-09-02

An environment or alcohol check-in. Awake duration is the sleep metric that responds most to late alcohol, room temperature, and noise. A tag-based feature that asks about the evening when awake time spikes above the user’s median gets useful answers.

Limitations

Wearable only. Phone-only users have neither field.
Count and duration can come from different logic. Count is scoped to the main session; duration from totals covers the provider’s whole night. They will not always agree.
Sensitivity varies by device. Do not compare users across devices on either field.
Absence is not zero.

Sleep efficiency for awake time as a fraction of time in bed
Sleep latency for awake time before sleep onset
Sleep stages for what else a wearable adds
Sleep Score explained for the seven sleep factors
Data dictionary in the developer docs

References

Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals. Sleep. 2004;27(7):1255-1273. https://doi.org/10.1093/sleep/27.7.1255