chore: instrument controlplane and cas with OpenTelemetry tracing

[javirln]

This pull request adds OpenTelemetry tracing support to the artifact-cas service, making tracing configurable via the application config and wiring it into the application's startup. The configuration schema, protobuf definitions, and generated code are updated to support tracing options. The gRPC server is now instrumented for tracing, and the configuration file includes a section for observability tracing settings.

Observability and Tracing Integration:

• Added a new observability.tracing section to the config schema (conf.proto and config.devel.yaml), allowing tracing to be enabled/disabled, endpoint configuration, insecure mode, and sampling ratio. [1] [2]
• Updated the protobuf definitions and generated Go code to include the new tracing configuration options (Bootstrap_Observability_Tracing message and related getters/setters in conf.pb.go). [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

Application Wiring and Startup:

• Modified the application wiring (wire.go, wire_gen.go, and main.go) to pass the full bootstrap config and initialize a TracerProvider if tracing is enabled. The tracer provider is now passed to the app and properly cleaned up on shutdown. [1] [2] [3] [4] [5] [6]

gRPC Server Instrumentation:

• Instrumented the gRPC server with OpenTelemetry by adding the otelgrpc stats handler, enabling distributed tracing for all gRPC requests. [1] [2]

Other:

• Updated copyright year in wire.go.

These changes make the service ready for distributed tracing with OpenTelemetry, improving observability and making it easier to diagnose issues in production environments.

[cubic-dev-ai]

5 issues found across 98 files

Note: This PR contains a large number of files. cubic only reviews up to 75 files per PR, so some files may not have been reviewed. cubic prioritises the most important files to review.

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="app/controlplane/pkg/biz/casbackend_checker.go">

<violation number="1" location="app/controlplane/pkg/biz/casbackend_checker.go:67">
P2: The new span wraps the entire checker lifecycle instead of each backend-check run, creating a long-lived parent span and coupling sampling across all periodic executions.</violation>
</file>

<file name="app/controlplane/internal/server/otel.go">

<violation number="1" location="app/controlplane/internal/server/otel.go:78">
P1: `sampling_ratio=0` is handled as `AlwaysSample`, so a config intended to disable tracing instead samples every trace.</violation>
</file>

<file name="app/controlplane/internal/conf/controlplane/config/v1/conf.proto">

<violation number="1" location="app/controlplane/internal/conf/controlplane/config/v1/conf.proto:66">
P2: `sampling_ratio` should use explicit presence (`optional`) to distinguish unset from an explicit `0.0` value.</violation>
</file>

<file name="app/artifact-cas/internal/server/otel.go">

<violation number="1" location="app/artifact-cas/internal/server/otel.go:77">
P2: Out-of-range `sampling_ratio` values are silently mapped to `AlwaysSample`, which can unexpectedly enable 100% tracing under misconfiguration.</violation>
</file>

<file name="app/controlplane/pkg/data/data.go">

<violation number="1" location="app/controlplane/pkg/data/data.go:129">
P2: `min_open_conns` is being mapped to `SetMaxIdleConns`, which does not enforce a minimum open pool size and changes config semantics.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review, or fix all with cubic.}

[jiparis]

Looks good. thanks!

[cubic-dev-ai]

2 issues found across 2 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="pkg/otelx/otelx_test.go">

<violation number="1" location="pkg/otelx/otelx_test.go:67">
P2: Do not reset the global OpenTelemetry tracer provider to nil in test cleanup; it can cause panics in subsequent tracer creation.</violation>
</file>

<file name="pkg/otelx/otelx.go">

<violation number="1" location="pkg/otelx/otelx.go:102">
P1: Returning `trace.SpanFromContext(ctx)` for disabled layers can end the parent span when callers `defer span.End()`. Return an explicit noop span instead.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review, or fix all with cubic.}