Epic: Implementing Metrics

[SergeyKanzhelev]

Implementing metrics falls into 4 buckets. API, instrumentation adapters, data pipeline, export. Breakdown is below.

For all the items - before starting, please create an issue and link it from this epic.

API implementation

• Basic API. Done by @cijothomas
• Handles and other optimizations
• API documentation
• API usability review

API Usage (instrumentation adapters)

• Expose metrics from instrumentation Adapter.
  • Record stats in HttpClient listener #68
  • Other? SQL, GRPC?
• Performance counters collection
  • Windows Performance Counters
  • Event Counters in .NET Core
  • Others

Data pipeline

• Ensure specs are in place
• Implement Metric Processor
• Implement metrics aggregations

Export

• Metrics Data Model. Define all the aggregates that SDK allows to configure out of the box.
• Prometheus exporter. See also https://github.com/open-telemetry/opentelemetry-dotnet/labels/area%3A%20Exporter.Prometheus
• OpenTelemetry Collector exporter