Filtering trace events by region

It's often useful to filter trace events on certain axes, such as those trace events generated by certain roles. Every trace event currently logs which roles are being run on the process to make this easier.

Another potentially useful filtering criterion in multi-region configs is which region the process is running in. For example, it can be helpful to see metrics specifically on the primary, satellite, or remote transaction logs.

There are technically at least two different pieces of information that are interesting here -- what region the processes are in and in what multi-region capacity (primary, remote, satellite). It's possible a solution could add one or both of these.

We should consider how to enable these types of queries without making logs overly verbose.

No files or tests are identified in the issue. Start by tracing how events currently record process roles and how multi-region processes expose region and capacity information. Done requires an agreed design for one or both filters, with logging verbosity considered and implementation coverage demonstrated by tests.

Written by the indexing model from the issue text.