Skip to main content
A Run is a defined time window during which telemetry is captured from one or more Assets. Runs provide the primary time context for organizing and analyzing telemetry.

​
Concurrent Runs

Multiple Runs of the same Asset can be active simultaneously. Each data point is owned by the Run that recorded it, so concurrent Runs on the same Asset do not collide even when they write to the same channel name at the same timestamp. This is the recommended pattern for parallel test sessions or CI workloads:
  1. Create one Asset for the system under test (for example, e1-sim).
  2. Start one Run per session, branch, or job. Name or tag each Run to make it identifiable later. For example, with a git SHA or CI job ID.

​
Asset view behavior when no Run is selected

When viewing an Asset in Explore without selecting a specific Run, Sift surfaces the most recently received value across all Runs. To isolate the data from a particular session, select that Run explicitly from the Run picker.

​
Runs and ad hoc Runs

An ad hoc Run is a Run created from data already in Sift, using the Create Run action on a Time Series Panel in an Asset’s or a Run’s Explore view. It records a name, time range, and one or more Assets, and includes all data on those Assets within that time range, regardless of which Run, if any, originally produced it.

​
File attachments

Files can be attached to a Run as supplemental metadata. Attachments do not modify the Run or its telemetry.

​
Supported file types

​
Size limits

​
Update a file attachment

You can’t update the contents of an existing file attachment in place. To replace a file’s contents, remove the outdated file and attach the new one.

​
Common operations