feat: initial support for diagnostics_channel TracingChannel API

[Neha-Sharmaaa]

Problem in current implementation

Currently, monitoring Mongoose operations from APM tools requires monkey-patching which is brittle and can lead to performance regressions.

Objective

Enable APM tools (like OpenTelemetry) to instrument Mongoose without monkey-patching, by emitting tracing lifecycle events using diagnostics_channel.

Scope

Implement support for:

• Query.prototype.exec
• Aggregate.prototype.exec
• Model.prototype.save

Features

• Follows the tracing lifecycle flow: start → operation → asyncEnd (with error on failure).
• Zero overhead when no subscribers are active.
• Contextual objects for each operation including filters, pipelines, and collection names.
• Compatible with Node.js 18+ and handles Node 20.x stable branch APIs correctly.

Note: This PR is now correctly separated from the populate-limit optimization PR.

[Neha-Sharmaaa]

Usage Example and Verification

You can subscribe to these channels to monitor Mongoose operations without monkey-patching:

const dc = require('node:diagnostics_channel');

// Monitor all Mongoose queries
dc.tracingChannel('mongoose:query').subscribe({
  start(ctx) {
    console.log('Query Start:', ctx.operation, 'on', ctx.collection);
    console.log('Filter:', ctx.query);
  },
  asyncEnd(ctx) {
    console.log('Query End');
  },
  error(ctx) {
    console.error('Query Error:', ctx.error);
  }
});

// Works similarly for 'mongoose:aggregate' and 'mongoose:save'

Verified locally with a new test suite (test/diagnostics_channel.test.js) in Node v20.19.4.

[Neha-Sharmaaa]

If CI checks aren't showing, a maintainer may need to approve workflow runs from this fork: go to the Actions tab, open the pending workflow run for this PR, and click Approve and run.