When adding more startup measurement tasks, we populated this sheet with the measurement times. We also moved to a new variant that we ran the task under. Doing so had shown our startup times drastically change from ~500s startup time for 50k collections to ~1200s for 50k collections.

We should run the same task on the old variant to ensure that this difference in time is because of the variant change, and not some code change that introduced a regression.

The previous variant was `Disagg 2-Node ReplSet ARM (mixed instances, Mongotune Dev Policies)`. The current variant is `Disagg 2-Node ReplSet Availability ARM (standby, Mongotune Dev Policies)`