Add SystemPerformanceInfo for compute platform telemetry

[follesoe]

Summary

Adds comprehensive system performance telemetry to support monitoring across both iMX6 (X3) and Jetson Orin NX 16 GB (X3 Ultra / X7) platforms — similar to what jtop, htop, and tegrastats provide.

New messages in message_formats.proto

Message	Purpose
CpuCoreLoad	Per-core CPU index, load (0..1), and clock frequency (MHz)
GpuInfo	GPU load and frequency
DlaInfo	Per-engine Deep Learning Accelerator load, frequency, and enabled state
MemoryInfo	RAM total/used/cached (uint64 to handle 16 GB+)
ThermalZone	Typed thermal zone reading using ThermalZoneId enum (TJ, CANISTER)
VideoCodecInfo	NVENC/NVDEC encoder and decoder active status and frequency
SystemPerformanceInfo	Composite message combining all of the above plus queue loads

PowerRailInfo — removed; INA3221 driver loaded but no devices bound on current carrier board.

New in telemetry.proto

• SystemPerformanceInfoTel — telemetry wrapper for SystemPerformanceInfo

Deprecations

• CPUInfo — superseded by SystemPerformanceInfo
• CPUTemperature — superseded by SystemPerformanceInfo.thermal_zones

Both are kept intact for backward compatibility.

Wire size estimates

Jetson Orin NX (all fields populated: 8 cores, GPU, 2 DLA engines, 2 thermal zones):

Field	Bytes
cpu_cores (x8)	~112
cpu_load_average	5
gpu	~17
dla_engines (x2)	~28
memory	~57
thermal_zones (x2)	~18
video_codec	~12
Queue loads (x6)	30
Total	~279 B

iMX6 (4 cores, 1 thermal zone, no GPU/DLA/codec): ~133 B

At 1-10 Hz publish rates this is well under 5 KB/s on the wire.

Design decisions

• ThermalZoneId enum instead of string for thermal zones — saves ~25 bytes and provides type safety. Currently defines TJ (junction temp) and CANISTER; more zones can be added as needed.
• uint64 for memory fields — uint32 would overflow at 4 GB, insufficient for the 16 GB Orin NX.
• Platform-agnostic composite message — unpopulated fields are zero/empty by default in protobuf, so iMX6 simply omits GPU/DLA/codec fields with no overhead.

[Copilot]

Pull request overview

Adds a new composite telemetry schema for system performance monitoring across compute platforms (iMX6 and Jetson Orin NX), introducing a richer replacement for legacy CPU-only telemetry while keeping backward compatibility.

Changes:

• Added SystemPerformanceInfo and supporting messages/enums to message_formats.proto (CPU cores, GPU/DLA, memory, thermals, power rails, video codec, queue loads).
• Added SystemPerformanceInfoTel wrapper to telemetry.proto.
• Marked CPUInfo / CPUTemperature and CPUInfoTel as deprecated via comments pointing to the new message(s).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
protobuf_definitions/telemetry.proto	Adds SystemPerformanceInfoTel and marks CPUInfoTel as deprecated in comments.
protobuf_definitions/message_formats.proto	Introduces SystemPerformanceInfo and related component messages; annotates legacy CPU messages as deprecated.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[jp-pino]

For reference this is the info we can probably get:

Probably also good to report the current power mode

[jp-pino]

Latest changes summary

DlaInfo

• Added bool enabled = 4; — reflects clk_enable_count >= 1 from debugfs clock tree (same approach as rbonghi/jetson_stats)
• Frequency now documented as "0 when disabled"

PowerRailInfo removal

• Removed PowerRailInfo message entirely
• Used reserved 7; in SystemPerformanceInfo to prevent field number reuse
• Rationale: Investigated all available sysfs power paths on Orin NX:
  • /sys/class/nvidia-gpu-power/ — no power metrics, just device class
  • /sys/class/nvidia-gpu-v2-power/ — same GPU device (load/freq only)
  • /sys/class/power_supply/ — empty
  • /sys/class/regulator/ — only configured voltage set-points, requested_microamps always 0
  • No INA current-sense hardware bound on carrier board

New queue load fields

• Added camera_queue_load = 12, overlay_queue_load = 13, position_observer_queue_load = 14
• All 6 EventsQueueWithStats instances now mapped to telemetry

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 8 comments.

Comments suppressed due to low confidence (1)

protobuf_definitions/message_formats.proto:1294

• The CPUInfo header says it includes "memory usage", but the message only reports memory_bus_load (plus CPU + queue loads). Consider adjusting the description to "memory bus load" (or add actual RAM/swap usage fields if that’s what consumers expect) so the deprecation guidance is clear.

// CPU information (deprecated, use SystemPerformanceInfo instead).
//
// Contains information about the CPU load and memory usage of the drone.
message CPUInfo {
  float cpu_load = 1; // CPU load (0..1).
  float memory_bus_load = 2; // Memory bus load (0..1).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (2)

protobuf_definitions/message_formats.proto:1294

• CPUInfo is documented as containing "CPU load and memory usage", but the message only has cpu_load, memory_bus_load, and queue loads (no RAM usage fields). Since this comment was touched in this PR (deprecation note), please update the remaining text to match the actual fields (e.g., "CPU load, memory bus load, and queue loads").

// CPU information (deprecated, use SystemPerformanceInfo instead).
//
// Contains information about the CPU load and memory usage of the drone.
message CPUInfo {
  float cpu_load = 1; // CPU load (0..1).
  float memory_bus_load = 2; // Memory bus load (0..1).

.github/workflows/ci-dotnet.yaml:30

• The sanitization can produce an empty $SANITIZED (e.g., a branch name like _ becomes - then gets trimmed). That would make --version-suffix start with a dot (".123"), which is not a valid SemVer/NuGet pre-release label and will break dotnet pack. Add a fallback when $SANITIZED is empty (e.g., default to branch/ci) before writing the output.

        run: |
          BRANCH="${GITHUB_REF_NAME}"
          SANITIZED=$(echo "$BRANCH" | sed 's/[^a-zA-Z0-9]/-/g' | sed 's/--*/-/g' | sed 's/^-//;s/-$//' | tr '[:upper:]' '[:lower:]' | cut -c1-63)
          echo "sanitized-branch-name=$SANITIZED" >> "$GITHUB_OUTPUT"

      - name: Build project and generate NuGet package
        run: |
          dotnet pack Blueye.Protocol.Protobuf.csproj --version-suffix "${{ steps.branches.outputs.sanitized-branch-name }}.$GITHUB_RUN_NUMBER" -c Release -o out

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.