Generated banner for Doca Telemetry Traceback Application
Generated logo for Doca Telemetry Traceback Application

Doca Telemetry Traceback Application

NVIDIAOfficialCreated by NVIDIA

  • Skill
  • Tool integration
  • Cloud and DevOps
  • Apache-2.0
  • Medium risk

From the author

Use this skill when the operator is driving the DOCA Telemetry Traceback application to monitor telemetry data locally on a BlueField DPU or ConnectX NIC, detect long-term trends and threshold breaches with user-defined rules, and capture pre-event history plus active-event data to a log file. Trigger even when the user does not explicitly mention "DOCA Telemetry Traceback" — typical implicit phrasings include "monitor telemetry trends on the DPU without streaming all raw data", "log the samples before and after a threshold breach", "set up a rule that triggers when rx bytes/s exceeds 1M", "differential greater than rule on a data ID", "circular buffer of telemetry samples", or "traceback event log". Refuse and route elsewhere: the DOCA Telemetry Diag programming API and raw counter reads to doca-telemetry, DOCA install/bring-up to doca-setup, and version questions to doca-version.
— the description by NVIDIA

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The DOCA Telemetry Traceback application is a tool for monitoring telemetry data locally on BlueField DPUs or ConnectX NICs. It allows users to define rules for detecting trends and threshold breaches, capturing both pre-event history and active-event data into a log file. This skill provides guidance on configuring the application via JSON, compiling it from source, and executing it on hardware, while maintaining clear boundaries with other DOCA components like the Telemetry Diag API or installation procedures.

What it does

  • Configures telemetry monitoring rules and data IDs via JSON.
  • Supports compilation of the application from source code.
  • Executes the application on BlueField DPU or ConnectX NIC hardware.
  • Captures pre-event and active-event telemetry data to log files.
  • Provides troubleshooting steps for empty logs or configuration mismatches.

When to use it

  • When you need to monitor telemetry trends locally on a DPU without streaming raw data.
  • When you need to set up rules to trigger logs based on threshold breaches.
  • When you need to compile the Telemetry Traceback application from source.
  • When you are diagnosing issues like empty event logs or unsupported data IDs.

Advantages

  • Enables local telemetry processing and traceback logging on supported hardware.
  • Provides structured guidance for rule definition and application execution.
  • Maintains clear separation between application usage and low-level API programming.
  • Integrates with standard DOCA development workflows.

Usage examples

“How do I set up local telemetry monitoring for port_0_rx_bytes on my DPU?”
The agent provides the necessary JSON configuration structure and rule definitions to monitor the specified data ID and log history around threshold breaches.
“Why is my differential_greater_than rule not triggering in the event log?”
The agent explains the requirement for sample history depth and delta logic, guiding you to verify your configuration parameters.

Install

  1. Download the skill as a ZIP:

    https://autogineer.com/agent-skills/nvidia/doca-telemetry-traceback-application/download
  2. Unzip it into the cross-client skills folder of your project:

    mkdir -p .agents/skills && unzip doca-telemetry-traceback-application.zip -d .agents/skills

Documentation by NVIDIA

DOCA Telemetry Traceback Application

Scope boundaries

These requests are out of scope: say so, name the skill that owns the request, and stop.

  • the DOCA Telemetry Diag programming API and raw counter reads: doca-telemetry/SKILL.md.

  • DOCA install and bring-up: doca-setup/SKILL.md.

  • which DOCA version is installed, version consistency: doca-version/SKILL.md.

  • general DOCA orientation and documentation lookup: doca-public-knowledge-map/SKILL.md.

Examples

  • “Can I monitor telemetry trends on the DPU without streaming all raw data?” — in scope. Configure rules and sample history depth; do not use the DOCA Telemetry Diag API directly here.

  • “The event log is empty even though a threshold was crossed.” — check rule comparison_type and sample_history_depth; differential rules require depth samples before functioning.

  • “How do I build only this application?” — use the Meson isolation flags; see the tasks companion.

  • “I'm writing my own C code against the DOCA Telemetry Diag library to read counters.” — that part is doca-telemetry's: the application uses Diag internally, but programming the Diag API is not this skill. Name doca-telemetry for it and do not write Diag API code here; handle every other part of the request as this skill describes (for example, --force-ownership still goes through doca-hardware-safety).

Quick facts

The authoritative source for the binary name, flags, and paths is the public DOCA Telemetry Traceback guide for the release; confirm them with the installed tool's --help before quoting any of them. As that guide describes it:

  • Standalone application on a BlueField DPU or ConnectX NIC; four components: DOCA Telemetry Diag instance, sample history circular buffer, rule collection, and events log file.

  • Configuration JSON fields: diag_sample_period_ns, sample_history_depth, data-ids array, and rules array. Rule types are incremental_greater_than (inc_gt), differential_greater_than (dt_gt), differential_less_than (dt_lt), greater_than (gt), and less_than (lt).

  • The circular buffer size is automatically doubled based on the requested depth to ensure equal historical and active event data.

  • Requires sudo/root privileges; mandatory -d PCIe address and -r rules JSON; optional --event-log and --force-ownership.

  • Unsupported data IDs are reported; adjust inputs accordingly.

  • The application stops if the actual sample size or period does not match the requested parameters.

  • --force-ownership changes diag domain ownership; route any use of it through doca-hardware-safety/SKILL.md.

Instructions

Follow Loading order below. Open the tasks companion for configure / build / run / debug. Open the capabilities companion for what the application monitors and logs. Do not open every companion file.

Where to start: This is a tool skill for invoking the DOCA Telemetry Traceback application — the local telemetry monitoring and traceback logging service that processes telemetry data on the device and writes event logs when user-defined rules trigger. Open the tasks companion and start at ## configure for the JSON configuration and rule definitions, then ## run for compile and execution. Open the capabilities companion when the question is what data can be monitored, what rules evaluate, and what the event log contains. If the user has not installed DOCA yet, route to doca-setup/SKILL.md first. If the user asks about version, route to doca-version/SKILL.md. If any step changes hardware or firmware state (e.g., --force-ownership), route to doca-hardware-safety/SKILL.md.

How to use this skill. Read this file, then search the companion file for the ## <name> heading the question needs and read that whole section. Answer completely from it, quoting only its symbols and commands. Do not browse docs; use the entry points in doca-public-knowledge-map. Give environment checks as commands for the user's DOCA host. For out-of-scope requests, name the skill that owns them and stop.

Compile and run commands are documented in the tasks companion; the authoritative source is the public guide. The application is distributed as source and must be compiled before execution. The source is located at /opt/mellanox/doca/applications/telemetry/traceback/.

Example questions this skill answers well

The CLASSES of telemetry-traceback questions this skill is built to answer, each with one worked example. The class is the load-bearing piece; the worked example is one instance.

  • “How do I set up local telemetry monitoring with traceback logging?” — worked example: “I want to monitor port_0_rx_bytes on the DPU and log history before and after it exceeds 1 MB/s”. Answered by the JSON configuration and rule definitions in the tasks companion.

  • “What rule types are available and when do they activate/deactivate?” — worked example: “I need a rule that triggers each time a counter increases by a threshold, not just when it exceeds a fixed value”. Answered by the rule definitions in the capabilities companion.

  • “Why is my differential rule not triggering?” — worked example: “I configured differential_greater_than but the event log stays empty”. Answered by the depth requirement and delta logic in the capabilities companion.

  • “How do I compile and run only this application?” — worked example: “I want to build just the Telemetry Traceback app, not all DOCA applications”. Answered by the Meson isolation flags in the tasks companion.

  • “What does the event log contain when multiple events overlap?” — worked example: “Two rules trigger at the same time; how are begin/end markers written?”. Answered by the event log lifecycle in the capabilities companion.

  • “How do I know if my device supports the data IDs I configured?” — worked example: “The application reports unsupported IDs; what do I do?”. Answered by the unsupported data ID handling in the tasks companion.

When to load this skill

Load this skill when the user is — or the agent needs to — drive the DOCA Telemetry Traceback application on a real host or BlueField Arm with DOCA installed (or inside the public NGC DOCA container with the right device passthrough). Concretely:

  • Configuring JSON rules and data IDs for local telemetry monitoring.

  • Compiling the application from source.

  • Running the application with a PCIe address and rules file.

  • Diagnosing an empty event log, unsupported data IDs, or sample size/period mismatch.

  • Understanding rule activation/deactivation semantics.

  • Cross-checking event log output against configured rules.

Do not load this skill for general DOCA orientation, the DOCA Telemetry Diag programming API, raw telemetry collection, DOCA install, or version questions. For those, route to doca-public-knowledge-map/SKILL.md, doca-setup/SKILL.md, or doca-version/SKILL.md.

What this skill provides

This is a thin loader. Substantive material lives in two companion files:

  • the capabilities companion — what the application monitors and logs: the four components, configuration schema, rule definitions with activation/deactivation triggers, event log lifecycle, unsupported data ID behavior, version overlay that redirects to doca-version/SKILL.md, and the safety policy for --force-ownership.

  • the tasks companion — step-by-step workflows for the in-scope task verbs: configure (JSON config, rule selection), build (Meson compile all or only this app), run (execution syntax, root, PCIe address), debug (empty log, unsupported IDs, sample mismatch), plus a Command appendix that honors the bundle's doca-structured-tools-contract/SKILL.md preamble.

The skill assumes a host or BlueField where DOCA is already installed (or the public NGC DOCA container is running with the right device passthrough) and the operator has root privileges as required by the public guide.

What this skill deliberately does not ship

This skill is agent guidance, not a samples or scripts bundle. To keep the boundary clean, it deliberately does not contain — and pull requests should not add:

  • Verbatim flag inventories, subcommand names, or output column names. The public DOCA Telemetry Traceback guide and the installed --help on the user's version are the joint source of truth; copying them here pins the skill to one release and silently rots when the application evolves. The skill routes the agent at those sources instead.

  • Pre-baked example output. Output is install-, version-, and device-specific. A captured example will mislead an operator on a different platform / state.

  • Wrappers, parsers, or scripts in any language that consume the application's output. The output format is documented; users who want to script against it should read the live guide and write the parser against their installed version.

  • A samples/ or reference/ subtree. This is a thin loader for a documented DOCA artifact; substantive material lives on the public page and in --help.

Loading order

  1. Read this skill first to confirm the user's question is in scope (the user wants to monitor telemetry data locally and capture traceback event logs, not use the DOCA Telemetry Diag programming API directly).

  2. For what the application monitors, the configuration schema, rule definitions, event log lifecycle, unsupported data ID behavior, version availability, and safety posture, see the capabilities companion.

  3. For the documented invocations and workflows — configure, build, run, debug, plus the Command appendix — see the tasks companion.

  • doca-hardware-safety/SKILL.md — the bundle's hardware/firmware state-change gate. The Telemetry Traceback application is read-only for telemetry data, but the --force-ownership flag changes diag domain ownership; route any use of that flag through this skill before execution.

  • doca-version/SKILL.md — canonical DOCA version-handling rules. The ## Version compatibility section in the capabilities companion is a concise overlay that redirects here for the body.

  • doca-setup/SKILL.md — env preparation, install verification, and the I have no install yet path with the public NGC DOCA container. This skill assumes its preconditions are satisfied.

  • doca-public-knowledge-map/SKILL.md — routing to the public DOCA Telemetry Traceback guide and the rest of the public DOCA documentation set. The canonical URL is reached via doca-public-knowledge-map ## DOCA tools (doca-public-knowledge-map/SKILL.md#doca-tools).

  • doca-structured-tools-contract/SKILL.md — the bundle's detect → prefer → fall back → report contract for structured helper tools. The Command appendix in the tasks companion honors this contract.

Files

BENCHMARK.md

Skill Benchmark: doca-telemetry-traceback-application

✅ Overall verdict: PASS — Recommended for publication

Publication Recommendation

Recommended for publication based on the completed evaluation evidence in this report.

Evaluation Metadata

  • Skill: doca-telemetry-traceback-application

  • Validation status: passed

  • Evaluation date: 2026-10-28

  • Evaluator version: 1.6.2

  • Agents: Claude Code (aws/anthropic/bedrock-claude-opus-5-5), Codex (openai/openai/gpt-6.1-sol)

  • Tasks: 5 evaluation tasks (3 positive, 2 negative)

  • Dataset digest: sha256:0c9167216e7c5952361cf854921a86d3854342935e14e647a42af4d35ac4adb4 (skill-evaluator-dataset-snapshot/1)

  • Attempts per task: 1

  • Environment: k8s-sandbox

  • Tier 2 evidence: required for publication

  • Tier 3 evidence: required for publication

Each task attempt ran in its own isolated sandbox pod.

Tier 1 LLM Provenance

SkillSpector provider, model, analyzer revision, and determinism evidence retained for this scan.

SkillProviderAdapterAnalyzerModelRevisionDeterminism
doca-telemetry-traceback-applicationnv_inferenceopenaimcp_tool_poisoningaws/anthropic/bedrock-claude-opus-5-52.13.1nondeterministic / provider_defaults
doca-telemetry-traceback-applicationnv_inferenceopenaisemantic_developer_intentaws/anthropic/bedrock-claude-opus-5-52.13.1nondeterministic / provider_defaults
doca-telemetry-traceback-applicationnv_inferenceopenaisemantic_quality_policyaws/anthropic/bedrock-claude-opus-5-52.13.1nondeterministic / provider_defaults
doca-telemetry-traceback-applicationnv_inferenceopenaisemantic_security_discoveryaws/anthropic/bedrock-claude-opus-5-52.13.1nondeterministic / provider_defaults
doca-telemetry-traceback-applicationnv_inferenceopenaimeta_analyzeraws/anthropic/bedrock-claude-opus-5-52.13.1nondeterministic / provider_defaults
  • Controls for doca-telemetry-traceback-application: temperature=not set (source=provider_default, forwarded=no); seed=not set (source=unset, forwarded=no); reasoning_effort=not set (source=provider_default, forwarded=no)

What This Report Answers

The three-tier evaluation checks whether the skill:

  • is safe to use;

  • produces correct answers;

  • is discovered and activated when needed;

  • helps the agent complete the user's goal and expected workflow; and

  • avoids wasted skill and tool usage.

Results at a Glance

MeasureClaude Code (Baseline → Skill Uplift)Codex (Baseline → Skill Uplift)
Overall96.0% — baseline ran, but no comparable score was available; uplift unavailable91.7% — baseline ran, but no comparable score was available; uplift unavailable
Security100.0% → 100.0% (±0.0 points)90.0% → 100.0% (+10.0 points)
Correctness60.0% → 100.0% (+40.0 points)56.0% → 92.0% (+36.0 points)
Discoverability100.0% — baseline ran, but no comparable score was available; uplift unavailable95.0% — baseline ran, but no comparable score was available; uplift unavailable
Effectiveness50.0% → 90.0% (+40.0 points)45.0% → 95.0% (+50.0 points)
Efficiency90.2% — baseline ran, but no comparable score was available; uplift unavailable76.3% — baseline ran, but no comparable score was available; uplift unavailable

How to read this table: baseline is the same task attempted without the target skill. Scores are rounded to one decimal; threshold-adjacent values use additional precision so their displayed band matches the verdict. Uplift is derived from those displayed scores and shown in percentage points.

Example: 47.0% → 92.0% (+45.0 points) means the skill-assisted run scored 92.0%, 45.0 percentage points above its 47.0% no-skill baseline.

A partial dimension was calculated from only the available configured signals; review the detailed report before relying on it.

Comparison details

Shared-source comparisons use the same signals and weights on both conditions. They do not replace the full dimension scores above or change the publication verdict.

AgentDimensionBasisBaseline → With skill (uplift)Explanation
claude-codeDiscoverabilityDimension comparisonN/ADiscoverability is not applicable to the no-skill baseline; there is no installed skill to discover.
claude-codeEfficiencyToken-only comparison98.2% → 94.3% (-3.9 points)Only Token Efficiency is comparable; the full dimension scores use different evaluator sources.
codexDiscoverabilityDimension comparisonN/ADiscoverability is not applicable to the no-skill baseline; there is no installed skill to discover.
codexEfficiencyToken-only comparison82.5% → 85.9% (+3.4 points)Only Token Efficiency is comparable; the full dimension scores use different evaluator sources.

Token Usage

Actual Tier 3 execution usage is reported for every observed agent/case pair and both conditions.

AgentDataset caseWith skillWithout skillDeltaChangeCoverage
claude-codeAll cases689,208467,900+221,308+47.30%skill 5/5; base 5/5
claude-codedoca-telemetry-traceback-application.explicit-positive.v1170,80671,605+99,201+138.54%skill 1/1; base 1/1
claude-codedoca-telemetry-traceback-application.implicit-positive.v1201,290107,791+93,499+86.74%skill 1/1; base 1/1
claude-codedoca-telemetry-traceback-application.implicit-positive.v2204,89171,301+133,590+187.36%skill 1/1; base 1/1
claude-codedoca-telemetry-traceback-application.negative.v136,663107,714-71,051-65.96%skill 1/1; base 1/1
claude-codedoca-telemetry-traceback-application.negative.v275,558109,489-33,931-30.99%skill 1/1; base 1/1
codexAll cases380,602370,166+10,436+2.82%skill 5/5; base 5/5
codexdoca-telemetry-traceback-application.explicit-positive.v193,74556,490+37,255+65.95%skill 1/1; base 1/1
codexdoca-telemetry-traceback-application.implicit-positive.v198,55822,459+76,099+338.84%skill 1/1; base 1/1
codexdoca-telemetry-traceback-application.implicit-positive.v292,57817,365+75,213+433.13%skill 1/1; base 1/1
codexdoca-telemetry-traceback-application.negative.v129,95339,429-9,476-24.03%skill 1/1; base 1/1
codexdoca-telemetry-traceback-application.negative.v265,768234,423-168,655-71.94%skill 1/1; base 1/1
ALL AGENTSDataset aggregate1,069,810838,066+231,744+27.65%skill 10/10; base 10/10

Prompt tokens include cached reads, so total tokens are prompt + completion (cached is not added twice). The Efficiency score uses (prompt - cached) + completion. Change is 100 * (with - without) / without; negative means fewer tokens. N/A comparisons include their reason in the coverage column; coverage is never estimated.

Tier Status

TierPurposeStatusEvidence
Tier 1Static validationPASSED WITH OBSERVATIONS11 validator(s); 23 finding(s)
Tier 2Semantic deduplicationPASSED WITH OBSERVATIONS2 validator(s); 3 finding(s)
Tier 3Live agent evaluationPASS2 agent(s); 5 task(s)

Findings and Observations

<details> <summary>Show detailed findings and successful checks</summary>

  • CRITICAL CONTENT_DEDUP/llm_error: LLM analysis failed for a content cluster (skills/tools/doca-telemetry-traceback-application)

  • HIGH DUPLICATE/duplicate: Duplicate content found within CAPABILITIES.md: "## Capabilities and modes" in CAPABILITIES.md (lines 106-113) vs "## Observability" in CAPABILITIES.md (lines 161-187) (CAPABILITIES.md:106)

  • HIGH DUPLICATE/duplicate: Duplicate content found across CAPABILITIES.md and TASKS.md: "## Error taxonomy" in CAPABILITIES.md (lines 148-160) vs "## debug" in TASKS.md (lines 292-312) (CAPABILITIES.md:148)

  • MEDIUM SCHEMA/metadata_key_style: Metadata key 'doca_version' is not kebab-case (skills/tools/doca-telemetry-traceback-application/SKILL.md)

  • MEDIUM SCHEMA/folder_hierarchy: Unexpected nesting depth for general skill (skills/tools/doca-telemetry-traceback-application)

  • 21 additional finding(s) are available in the full evaluation artifacts.

</details>

Scoring Methodology

<details> <summary>Show dimension definitions, source signals, and thresholds</summary>

DimensionQuestionScored signals
SecurityIs it safe to use?security (100%)
CorrectnessIs the answer correct?accuracy (100%)
DiscoverabilityWas the right skill loaded when needed?skill_execution (100%)
EffectivenessDid the skill help complete the task?goal_accuracy (50%) + behavior_check (50%)
EfficiencyDid it avoid wasted tool calls and token usage?skill_efficiency (50%) + token_efficiency (50%)
  • Dimension bands: PASS at 50% or above; NEUTRAL from 40% to below 50%; FAIL below 40%.

  • Overall Tier 3 lift: PASS at +5 points or more; FAIL at -10 points or less; values between those bands are NEUTRAL.

  • Overall verdict: PASS only when every configured dimension passes for at least one supported agent. Lift is reported as diagnostic evidence and does not override this gate.

  • The 50% attempt pass threshold is a separate per-task gate; it is not the dimension pass threshold.

  • Effectiveness is the equal-weight mean of goal completion (goal_accuracy) and expected workflow adherence (behavior_check).

  • Efficiency is 50% tool-call productivity (the backward-compatible skill_efficiency wire id) and 50% token_efficiency. Positive-case skill routing is scored under Discoverability, not Efficiency; a negative case without a routing target is N/A. N/A sources are omitted, remaining weights are renormalized, and the dimension is marked partial.

Signals present in this run:

  • security (Security): unsafe operations, secret leakage, and unauthorized access.

  • skill_execution (Skill Execution): whether the expected skill was selected, decoys were avoided, and the workflow executed.

  • skill_efficiency (Tool Productivity): tool-call productivity (legacy wire id; routing is scored under Discoverability).

  • accuracy (Accuracy): final-answer correctness against the reference answer.

  • goal_accuracy (Goal Accuracy): whether the user's goal was achieved.

  • behavior_check (Behavior Check): whether the expected workflow behavior was followed.

  • token_efficiency (Token Efficiency): actual uncached prompt plus completion usage (50% of Efficiency).

</details>

Freshness

Regenerate this benchmark when the skill, evaluation dataset, target agent/model, evaluator version, environment, or scoring policy changes.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

The author says: Requires DOCA SDK installed at /opt/mellanox/doca on Linux with a BlueField DPU or ConnectX NIC, plus root privileges to access hardware-level telemetry data. The application is distributed as source code and must be compiled before execution; the source is located at /opt/mellanox/doca/applications/telemetry/traceback/. Validation steps require a live DOCA install.

Security

What our scanner noticed· Medium risk

  • Large base64-encoded blobs

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by
NVIDIA
Publisher
NVIDIA (Official)
Source
NVIDIA/skills · skills/doca-telemetry-traceback-application · commit fce3cca
Licence
Apache License 2.0 (redistribution allowed)
  • LICENSE (added by AutoGineer from LICENSE-APACHE)
  • ATTRIBUTION.md (written by AutoGineer)
Synced

AutoGineer is not affiliated with or endorsed by NVIDIA or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download doca-telemetry-traceback-application.zip