So the purpose of this continue is to skip the part at the end of the loop that waits for the next tick:

		select {
		case <-pingTicker.C:
			continue
		case <-a.stop:
			return nil
		}

And because we're effectively forcing a ping, pingTicker.Reset(pingInterval) sets the clock back to a full ping duration, so that the time since the last ping doesn't come out of the next iteration of the loop.

Makes sense 👍🏼

That's the intent! Hopefully a middle ground between good responsiveness without an increase in overall/average load placed upon Buildkite.