Add AWS Lambda worker support

[eamsden]

What was changed

• Added the experimental temporalio-sdk-aws-lambda integration crate.
• Added a reusable Lambda worker builder that creates a fresh Temporal client and Worker for each invocation.
• Added Lambda-oriented slot, poller, cache, eager Activity, versioning, and graceful-shutdown defaults, while allowing an explicit custom Worker tuner.
• Added client configuration loading from environment variables and Lambda-local temporal.toml files.
• Added absolute-deadline shutdown handling, per-invocation Worker identity, ordered shutdown hooks, tests, documentation, and changelog entries.
• Added optional Lambda OpenTelemetry setup for OTLP metrics and existing SDK spans, with ADOT-compatible endpoint and service-name defaults, AWS X-Ray-compatible trace IDs, and deadline-bounded per-invocation flushing.

Why?

AWS Lambda invocations have a bounded execution window and need time reserved for graceful Worker shutdown, telemetry export, and final cleanup. This adapter centralizes those host-specific lifecycle concerns and applies the constrained defaults used by the other Temporal SDK Lambda integrations. A general Rust SDK OpenTelemetry plugin, including cross-process trace-context propagation, is intentionally outside this PR.

Checklist

1. Closes: N/A

2. How was this tested:

• cargo test -p temporalio-sdk-aws-lambda
• cargo test -p temporalio-sdk-aws-lambda --features otel
• cargo lint
• cargo test-lint
• cargo +nightly fmt --check
• RUSTDOCFLAGS=-Dwarnings cargo doc -p temporalio-sdk-aws-lambda --features temporalio-sdk-aws-lambda/otel --no-deps
• git diff --check

An AWS Runtime Interface Emulator deployment smoke test remains as follow-up work.

3. Any docs updates needed?

The new crate includes a README covering setup, ADOT configuration, and custom-runtime guidance. Broader docs.temporal.io coverage and a deployable sample can follow once the experimental API is validated.