Improved metrics

[SVappsLAB]

Improved Metrics

• Adds System.Diagnostics.Metrics surface for telemetry throughput, drops, decode time, pipeline latency, handler time, session-info processing, and paced-IBT playback lag.

• Provides information on where data is lost: live missed ticks, async-consumer buffer eviction, session-info parse backlog, and public session-stream overflow.

• Adds metric tags for data source, drop reason, affected session-info stream, and processing errors, for monitoring dashboards and alerts.

• Provides per-stage latency decomposition

  acquire ──► decode ──► [ delivery buffer ] ──► handler start ──► handler end
    │         └ decode.duration ┘                  │                 │
    ├──────────── sdk.duration ──────────────────► │                 │
    │                                              ├ handler.duration ►
    └─────────────────── total.duration ────────────────────────────►
  

• Adds new Metrics samples. One using .Net's built-in MeterListener and one using OpenTelemetry for Grafana Dashboards

  Sample output from **MeterListener** - single stats sample, grabbed from console output:
  
  --- final metrics ---
   session_info.parse.attempts              1 recorded   avg          1 {attempt}   max          1 {attempt}
   session_info.process.duration            1 recorded   avg   62.944 ms   max   62.944 ms
   session_info.records.processed           1 {record}
   session_info.size                        1 recorded   avg     26,463 By   max     26,463 By
   telemetry.decode.duration          123,936 recorded   avg      0.5 us   max    1.696 ms
   telemetry.handler.duration         123,936 recorded   avg      0.2 us   max    206.7 us
   telemetry.records.processed        123,936 {record}
   telemetry.sample.interval          123,935 recorded   avg      2.3 us   max    3.349 ms
   telemetry.sdk.duration             123,936 recorded   avg      0.8 us   max    2.536 ms
   telemetry.total.duration           123,936 recorded   avg      1.2 us   max    2.715 ms