`dotnetup` needs to adopt the new telemetry exporter pattern

https://github.com/dotnet/sdk/issues/55184 This issue outlined core functionality gaps in Azure's OTel Exporter for standalone, one and done executables like CLI apps, which applies to `dotnetup`.

https://github.com/dotnet/sdk/pull/55211 Created its own system to manage telemetry until Azure accepts/rejects the applied API changes we requested.

We need to consider:
- The pattern that this PR uses to convert SDK events into events that go into the traces table and if that code is now duplicated across it and `dotnetup`
- If we need to adjust the CI timeouts in dotnetup or anything else on top of the changes inside of https://github.com/dotnet/sdk/pull/55155 with the changes in the dotnet/sdk PR
- Ensure that codeflow has merged in
- Adapt any further changes in telemetry here to use the new pattern
- Observe datum after this change to see if we're actually getting reliable data now or not and ensure there is not a perf regression outside of minimal CI waits to ensure flushing. Perhaps document the controllable flush backstop so users who wrote CI pipelines that do more after installing `dotnetup` in the same step don't need to wait on this.