Open an app and select Analytics to see usage alongside update diagnostics.
Dates and reporting periods
Choose From (UTC) and To (UTC), then select Apply. Both dates are included in full, up to 366 days. The default 7d range includes today and the preceding six UTC dates. Today means the current UTC calendar day, not the preceding 24 hours. Today's results are partial until that day ends.
The date selection controls App usage, Events in Range, Recent Events, and Export events CSV. The usage heading shows the dates currently loaded; editing the inputs does not change the report until you apply them. The CSV contains raw events, including updater events, rather than the daily usage tables. Its end date includes events through 23:59:59.999 UTC.
Other reports keep their stated periods: Events (30d) and Event Types cover the last 30 days, Active Devices (7d) covers the last seven days, and Deployment State and Top Errors cover the last 30 days. Channel adoption and total devices show the current registered-device state. These reports are not historical snapshots of the selected dates.
App usage
Usage requires the app to send session, page, or feature events. Update checks alone do not count as app usage. Older app versions without usage instrumentation will not contribute activity; there is no historical backfill.
| Metric | Meaning |
|---|---|
| Active installations | Unique app installation IDs with usage activity during the selected dates. One person can have multiple installations; shared installations can represent several people. This is not a count of identified people. |
| Sessions | Distinct installation and session ID pairs observed during the selected dates. |
| Average session time | Recorded foreground time divided by the session count. An em dash means no sessions were recorded. |
| Total foreground time | Time the instrumented app reports being in the foreground, excluding background time. It is not a measure of continuous interaction. |
| Page views | Recorded page-view events. |
| Feature events | Recorded named feature interactions. |
| Errors | Error events during the selected dates, reported separately from usage activity. |
| Installations with errors | Unique installations that reported errors during the selected dates. |
Foreground time comes from incremental session heartbeats and the session's final report. If the app terminates before sending its final report, the unsent interval is not counted. When a session crosses midnight UTC, daily rows describe that day's reported activity; a session may appear on more than one day. Range-level unique installations and sessions are deduplicated across the range, so adding daily unique counts does not produce the range total.
Daily activity charts one point per UTC date. Choose a chart metric to compare active installations, sessions, average session time, page views, feature events, or errors. Use touch, the mouse, or arrow keys on the chart to read individual days; Home and End move to the first and last day. Today may be incomplete. Expand Daily breakdown for all exact counts and durations, with sorting and pagination on desktop and readable day cards on smaller screens. Popular pages and Popular features rank recorded activity with relative bars, with the number of distinct installations for each item. A missing feature means no event was received for it; it does not prove nobody used it.
Empty tables indicate no recorded activity for the selected dates. A load failure is shown separately and can be retried with Apply. Switching apps or applying another range clears the old report while the new data loads.
Error diagnostics
The usage section's error totals follow the chosen UTC dates. Top Errors (last 30 days) groups recent errors by message, with affected devices, occurrences, versions, and the most recent occurrence. On phones, each error appears as a card so these details remain visible without horizontal scrolling. When the report is sampled, its heading states the sample size. Use both views to distinguish daily error volume from recurring error messages.