There are currently some server parameters related to the TTL Monitor, as can be found here. However, these only allow us to configure:

• How often the TTL Monitor wakes up to perform passes
• How much work is down within a TTL batched delete
  However, there is currently no way to limit how much work is done across an entire TTL Monitor pass. This can lead to behavior where the TTL monitor performs many passes that do no work, and then a few passes that perform a lot of work (this can be particularly evident in time-series collections where many bucket documents that were created around the same time can become eligible for deletion at the same time). This can lead to spikes in CPU usage.

Allowing the amount of total work that can be down in a TTL monitor pass could smooth out the deletion work across more passes, and lead to lower spikes.