https://findit-for-me.appspot.com/waterfall/flake?key=ag9zfmZpbmRpdC1mb3ItbWVy2QELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCKiAWNocm9taXVtLm1hYy9NYWMxMC4xMiBUZXN0cy81NS9icm93c2VyX3NpZGVfbmF2aWdhdGlvbl9icm93c2VyX3Rlc3RzIG9uIE1hYy0xMC4xMi9RMjl1ZEdsdWRXVlhhR1Z5WlVsTVpXWjBUMlptVkdWemRDNURiMjlyYVdWelEyeGxZWEpsWkU5dVEyeHZjMlZCYkd4Q2NtOTNjMlZ5Y3c9PQwLEhNNYXN0ZXJGbGFrZUFuYWx5c2lzGAEM

In this case, since there are so few build numbers, the earliest build number actually hits 1 so the regression range's lower bound stops there, however there are actually more commits before that

This is more of a corner case, but eventually we don't want to rely on build numbers necessarily