Time series collection for logs (high volume)

I’m working on an application which currently generates 100 - 600 MB of txt log files per hour (during busy hours, approaching 0 during night). Is it viable to switch to MongoDB time series collection for such an application, if one of the requirements is that the collection stores 6 month of data before archiving.
If it is, what kind of a setup would your recommend? Would a single VM with enough storage suffice, or would sharding be a must?