UnixMillisecond

Low-precision time is mostly used with persisted data structures and algorithms. It excels at logging and scheduling but it usually does not have sufficient precision to represent things such as network latencies or function execution times. It uses Milliseconds as its stride type.

It is worth noting that the stride durations are physically non-contiguous, as Unix time skips over leap seconds.

Why milliseconds?

Second-precision time is not always granular enough for logging, while microsecond-precision time limits the range of representable stride durations. Millisecond-precision time is a good compromise and also round-trips through BSON.