Add multi-instance SDK support via delegating providers

[JacksonWeber]

Summary

Adds opt-in support for running multiple isolated Microsoft OpenTelemetry SDK instances in a single Node.js runtime. Today useMicrosoftOpenTelemetry() registers global providers per signal, so a second initialization clobbers the first. This PR introduces a parent/child delegating-provider architecture so independent pipelines (e.g. two Application Insights resources with different connection strings) can coexist.

GA constraint respected: all changes are additive and non-breaking. The existing single-instance useMicrosoftOpenTelemetry() path is untouched.

What's new

• createMicrosoftOpenTelemetryInstance(options, { makeDefault? }) returns a MicrosoftOpenTelemetryInstance handle (getTracer/getMeter/getLogger, runWithInstance, forceFlush, shutdown).
• runWithMicrosoftOpenTelemetryInstance(id, fn) binds an ambient current instance so code using the global OTel API routes to the right pipeline.

Design

src/distro/multiInstance/:

• instanceRegistry.ts — registry + default instance + AsyncLocalStorage-backed ambient current instance.
• delegatingProviders.ts — Parent Tracer/Meter/Logger providers that resolve the current child per call (never cached) and delegate, with Noop fallbacks.
• globalSetup.ts — registers the parent providers + a single shared context manager + composite propagator exactly once.
• instance.ts — builds each child as standalone NodeTracerProvider/MeterProvider/LoggerProvider (no NodeSDK.start(), no global registration), wiring Azure Monitor handlers + caller processors + console fallback, then registers it.

Routing model: parents resolve a single current instance (ambient via runWithInstance → default fallback) and delegate, so telemetry stays isolated per instance rather than fanned out.

Testing

• New test/internal/functional/multiInstance.test.ts: two instances with distinct (fake) connection strings; asserts each instance's spans reach only its own exporter, and that global-API spans inside runWithInstance route to the bound instance.
• Verified live against two real Application Insights resources via a standalone sample app (all breeze /v2.1/track requests returned HTTP 200).
• Full unit + functional suites green; ESM + CJS build, lint, and format pass.

Dependencies

• Promotes @opentelemetry/context-async-hooks from transitive to a direct dependency (now imported directly to register the context manager, since the multi-instance path does not call NodeSDK.start()). No overrides used.

[Copilot]

Pull request overview

Adds an opt-in multi-instance Microsoft OpenTelemetry SDK mode that allows multiple isolated telemetry pipelines to coexist in a single Node.js process by registering global delegating providers that route per-call to an “ambient” (AsyncLocalStorage-bound) instance, falling back to a default instance.

Changes:

• Introduces createMicrosoftOpenTelemetryInstance() / runWithMicrosoftOpenTelemetryInstance() and the MicrosoftOpenTelemetryInstance handle type.
• Implements multi-instance runtime: instance registry + id binding, one-time global setup (context manager/propagator + parent providers), and delegating tracer/meter/logger providers.
• Adds functional coverage for isolation and ambient routing; promotes @opentelemetry/context-async-hooks to a direct dependency.

Reviewed changes

Copilot reviewed 10 out of 11 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
test/internal/functional/multiInstance.test.ts	New functional tests validating per-instance isolation and ambient routing via global API
src/types.ts	Adds MicrosoftOpenTelemetryInstance public interface type
src/index.ts	Re-exports new multi-instance APIs/types from the package entrypoint
src/distro/multiInstance/instanceRegistry.ts	Registry + AsyncLocalStorage-backed “current instance” binding and resolution
src/distro/multiInstance/instance.ts	Builds per-instance child providers/pipelines and lifecycle methods (flush/shutdown)
src/distro/multiInstance/index.ts	Exposes multi-instance APIs from the distro layer
src/distro/multiInstance/globalSetup.ts	One-time process-global setup for context manager/propagator + parent providers
src/distro/multiInstance/delegatingProviders.ts	Parent providers + delegating tracer/meter/logger implementations
src/distro/index.ts	Surfaces multi-instance exports through src/distro
package.json	Adds @opentelemetry/context-async-hooks direct dependency
package-lock.json	Locks the new dependency for npm ci installs

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[rads-1996]

Is it possible to break the PR into small chunks, so it is easier to review? Or does it make logical sense to have everything together?

[JacksonWeber]

Is it possible to break the PR into small chunks, so it is easier to review? Or does it make logical sense to have everything together?

I don't believe that will work well for this PR as it implements a single contained feature. If a single component of it is split off it won't function. Only way I can think to split it is per signal, but I'm not sure that'd make it much easier to review.