Add TimeseriesSessionCollector with memory and CPU collection

[barboraplasovska]

What and why?

Adds TimeseriesSessionCollector, the core sampling engine. Uses a DispatchSourceTimer at 1Hz to collect memory (phys_footprint via task_info) and CPU (Mach thread ticks summed across threads; nil on watchOS). Every 30-sample batch is flushed as a single event. A coin flip at session start decides whether the session emits the full array schema or the delta-encoded schema, so the backend receives both formats across the fleet.

Review checklist

• Feature or bugfix MUST have appropriate tests (unit, integration)
• Make sure each commit and the PR mention the Issue number or JIRA reference
• Add CHANGELOG entry for user facing changes
• Add Objective-C interface for public APIs - see our guidelines (internal)
• Run make api-surface when adding new APIs

[barboraplasovska]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• Wire timeseries collection into RUM session and configuration #2885
• Add TimeseriesSessionCollector with memory and CPU collection #2884 👈 (View in Graphite)
• Add delta encoder for timeseries compression #2883
• Generate timeseries memory and CPU RUM event models #2882
• Fix ObjC bridge generator for nested DataPoint types #2881
• feature/timeseries

This stack of pull requests is managed by Graphite. Learn more about stacking.