feat: add opt-in OpenTelemetry to chat turns and transcript loading

[calebeden]

Additional instructions

MUST: Keep Allow edits from maintainers enabled for this PR so maintainers
can help update the branch when needed.

What Problem This Solves

Native tray chat previously provided no OpenTelemetry visibility into a turn's
queueing, gateway admission, first-response latency, streaming duration, or
terminal outcome. Operators could see gateway diagnostics but could not explain
where time was spent from the Windows tray user's perspective.

Why This Change Was Made

This adds privacy-preserving chat lifecycle instrumentation with an explicit root
span per turn and sibling spans for send, response wait, and response receive.
Correlation identifiers remain process-local, exported values are finite and
low-cardinality, and prompts, responses, IDs, attachments, provider details, and
error messages are not exported.

Cross-process trace-context propagation into the gateway is intentionally left
for a separate coordinated gateway protocol change.

User Impact

Users who opt in to OpenTelemetry export can diagnose native chat latency and
outcomes in their existing collector without exposing conversation content.
Telemetry remains disabled when no endpoint is configured.

Evidence

The new tray spans and metrics were exercised through tracker and provider tests
covering local and remote turns, queueing and retries, first output, terminal
races, abort/reset/disconnect cleanup, bounded attributes, and privacy.

Screenshots

Change Type

• Bug fix
• Feature
• Refactor
• Docs or instructions
• Tests or validation
• Security hardening
• Chore or infrastructure

Scope

• Tray or WinUI UX
• Windows node capability
• Local MCP or winnode
• Gateway, connection, or pairing
• Setup or onboarding
• Permissions, privacy, or security
• Tests, CI, or docs

Validation

• .\build.ps1 - passed; all five projects built successfully.
• dotnet test .\tests\OpenClaw.Shared.Tests\OpenClaw.Shared.Tests.csproj --no-restore
  • 2,788 passed, 31 skipped, 0 failed.
• dotnet test .\tests\OpenClaw.Tray.Tests\OpenClaw.Tray.Tests.csproj --no-restore
  • 1,737 passed, 0 skipped, 0 failed.
• Focused ChatTelemetryTrackerTests and OpenClawChatDataProviderTests
  • 245 passed, 0 skipped, 0 failed.
• Final code review reported no significant issues.

Real Behavior Proof

• Environment tested: Windows ARM64 tray with an OpenTelemetry endpoint viewed
  through an Aspire dashboard.
• PR head or commit tested: def4fe81
• Exact steps or command run: launched the current tray build, sent a native chat
  message, and inspected the exported trace in Aspire.
• Evidence after fix: openclaw.chat.turn, openclaw.chat.send,
  openclaw.chat.response.wait, and openclaw.chat.response.receive were
  exported.
• Observed result: the tray trace decomposes admission, first-response wait, and
  response receive time without exported chat content or identifiers.
• Screenshot or artifact links verified? (Yes/No/N/A): Yes - screenshots included above
• Not verified or blocked: gateway openclaw.message.processed remains a separate
  trace because the current gateway protocol does not accept propagated W3C trace
  context.

Security Impact

• New permissions or capabilities? (Yes/No): No
• Secrets or tokens handling changed? (Yes/No): No
• New or changed network calls? (Yes/No): No
• Command or tool execution surface changed? (Yes/No): No
• Data access scope changed? (Yes/No): No
• If any answer is Yes, explain the risk and mitigation: N/A

Compatibility and Migration

• Backward compatible? (Yes/No): Yes
• Config or environment changes? (Yes/No): No
• Migration needed? (Yes/No): No
• If yes, list the exact upgrade steps: N/A

Review Conversations

• I replied to or resolved every bot review conversation addressed by this PR.
• I left unresolved only conversations that still need maintainer judgment.

[clawsweeper]

ClawSweeper status: review started.

I am starting a fresh review of this pull request: feat: add opt-in OpenTelemetry to chat turns and transcript loading This is item 1/1 in the current shard. Shard 0/1.

This placeholder means the worker is alive and reading the current context. I will edit this same comment with the actual review when the claws are done clicking.

Crustacean status: shell secured, claws on keyboard, evidence pebbles being sorted.