Problem
Looking at the PerfStar dashboard, several scenarios (e.g., .NET 8 console-app cold rebuild) only have data points every 4–7 days. This makes it hard to attribute regressions and improvements to specific MSBuild or SDK commits and slows our day-over-day feedback loop on the multithreading initiative.
Proposed change
Increase the PerfStar pipeline cadence to daily by default. Coordinate with the owners of the shared hardware before flipping the cadence — this team has previously received pushback when relay-bandwidth usage spiked during periods of active development, so social validation is a prerequisite.
Exit criteria
- PerfStar produces a data point for tracked scenarios at least once per day on working days.
- Pipeline runs are accepted by the shared-hardware owners as not creating undue load on shared infra.
Problem
Looking at the PerfStar dashboard, several scenarios (e.g., .NET 8 console-app cold rebuild) only have data points every 4–7 days. This makes it hard to attribute regressions and improvements to specific MSBuild or SDK commits and slows our day-over-day feedback loop on the multithreading initiative.
Proposed change
Increase the PerfStar pipeline cadence to daily by default. Coordinate with the owners of the shared hardware before flipping the cadence — this team has previously received pushback when relay-bandwidth usage spiked during periods of active development, so social validation is a prerequisite.
Exit criteria