elastic / cloudbeat / 16270040091

76%

Build Type

push

github

Committed by web-flow

Commit Message

Add OpenTelemetry (#3430)

### Summary of your changes
This pull request introduces OpenTelemetry (OTel) for observability, enabling tracing and metrics collection in cloudbeat.

The key changes are:
*   **New Observability Package:** A new package `internal/infra/observability` has been added to encapsulate all OpenTelemetry-related logic. This package is responsible for:
    *   Initializing and configuring the OTel tracer and meter providers.
    *   Setting up the OTLP exporter to send telemetry data to Elastic's APM server over gRPC.
    *   Providing helper functions to create spans and instrument code.
    *   Adding OTel instrumentation to AWS SDK calls.
*   **Instrumentation of Core Components:**
    *   The **Posture** flavor now initializes OpenTelemetry at startup and gracefully shuts it down on exit. This needs to happen for CNVM and Asset Inventory as well but it's left as a future item.
    *   The **Resource Fetching Manager** is instrumented to create spans for each fetching cycle and for each individual fetcher. This will help in tracking the performance and status of resource fetching.
    *   The **AWS Organization Benchmark** is updated to create spans for the AWS account discovery process. This required changes in the function signatures of the resource fetching `Registry`.
*   **Context Propagation:** The `context.Context` is now used to propagate trace information across different parts of the application.
*   **Enhanced Logging:** The logging infrastructure has been updated to include `trace.id` and `span.id` in the logs, allowing for easy correlation between logs and traces. More work is needed to fully utilize this across the codebase.

### Screenshot/Data

On AWS multi-account onboarding, the `otelaws` has an edge case where there's no span name when the ec2imds service cannot be reached:
```
Region 'us-east-1' selected after failure to retrieve aws regions: operation error ec2imds: GetInstanceIdentityDocument, canceled, co... (continued)

Run Details

165 of 224 new or added lines in 9 files covered. (73.66%)

2 existing lines in 1 file now uncovered.

9478 of 12450 relevant lines covered (76.13%)

16.6 hits per line