dt-obs-genai

AI Observability (GenAI) Skill

Analyze AI Observability signals from customer GenAI applications using DQL — golden signals, LLM signals, token and cost analytics (with usage attribution and prompt-caching economics), agent signals (including loop/runaway detection and Smartscape topology), conversation/session-level analytics, guardrails, and evaluation quality.

When to Use

Use this skill for observability questions about customer GenAI applications — anything reading OpenTelemetry GenAI spans (gen_ai.*) or LLM evaluation bizevents. Example triggers:

• "LLM latency", "error rate by model"
• "token usage by model", "token throughput / TPM", "am I hitting rate limits", "provider throttling or 429s"
• "cost by model and provider", "who is driving token spend", "do I have prompt caching"
• "cost per conversation", "most expensive sessions"
• "failing agent tool calls", "find runaway agents"
• "responses truncated or blocked", "finish reasons"
• "failed evaluations", "quality scores"

When NOT to Use

Davis CoPilot/MCP telemetry (dt-platform), generic service metrics (dt-obs-services), logs (dt-obs-logs), or non-GenAI distributed tracing (dt-obs-tracing).

Example Questions

When suggesting follow-up questions (e.g., "give me one example question per topic"), use these canonical, ready-to-ask phrasings — one per capability. Keep suggestions single-clause and avoid the literal phrase "content filter" (use "blocked or safety-filtered" instead); overly long, multi-clause questions can be rejected.

• Traffic, errors & latency: What is my LLM error rate and p95 latency by model in the last 24 hours?
• Token usage & cost: Break down token usage and cost by model and provider in the last 24 hours.
• Agent & tool activity: Show me failing agent tool calls in the last 24 hours.
• Conversation analytics: What are my top 10 most expensive conversations by total token usage in the last 24 hours?
• Guardrails: How many responses were blocked or truncated in the last 7 days, and which model was affected most?
• Evaluation quality: Show me failed evaluations from the last 24 hours with the judge's explanation and the question-answer pair.

Core Capabilities

Golden Signals

The four classic observability signals — traffic, errors, latency, and saturation — apply directly to GenAI applications. Traffic is request throughput over time; errors are spans where span.status_code == "error"; latency is the duration field (a Grail duration value — divide by the 1ms literal, duration / 1ms, for a numeric millisecond value); saturation is proxied by total token throughput per minute (input + output tokens combined).

fetch spans, from: now()-24h
| filter isNotNull(gen_ai.request.model)
| summarize total =

→ Full traffic, latency, and saturation queries: See references/golden-signals.md

LLM Signals

LLM signals describe which model and provider served each request, what operation type was invoked (chat, execute_tool, invoke_agent, create_agent), and how tokens were consumed. Use these to benchmark provider latency, compare model performance, and understand the token distribution across model-provider combinations.

→ Slowest models, token usage by model: See references/llm-signals.md

Cost and Tokens

Token consumption is the primary cost driver. Dynatrace stores gen_ai.usage.input_tokens and gen_ai.usage.output_tokens on every span — there is no stored cost field; estimated cost must be derived by multiplying token sums by the per-model price you supply. Use these queries to identify the highest-spend model-provider combinations and detect token-burn spikes.

→ Token spikes, cost estimation, most expensive prompts, usage attribution, prompt-caching economics: See references/cost-and-tokens.md

Agent Signals

GenAI agents emit spans for each tool invocation (execute_tool), agent step (invoke_agent), and agent creation (create_agent). Use agent signals to identify which tools are called most often and which agents are failing. For structural questions — which agents, models, and providers exist and how they connect — query the GenAI Smartscape entities (GENAI_AGENT, GENAI_MODEL, GENAI_PROVIDER, GENAI_SERVICE) instead of scanning spans; this is a feature-flag-gated preview.

→ Tool usage, failing agents, agent step latency, loop/runaway detection, Smartscape topology: See references/agent-signals.md

Conversation Analytics

Per-span and per-trace signals measure one request or one turn. When the application propagates gen_ai.conversation.id, you can roll spans up to the session level — cost per conversation, how deep conversations run, and which sessions are runaway-expensive or error-prone. This is the unit that matters for chargeback and user-perceived reliability.

→ Cost/depth per conversation, session error rate: See references/conversation-analytics.md

Guardrails

Guardrails surface as gen_ai.response.finish_reasons on the span — content_filter means a safety filter blocked or redacted output, length means the response was truncated at the token limit — and as the proactive safety evaluators (prompt-injection, pii-leakage, toxicity, bias) in the evaluation bizevents. Use these to quantify blocked and truncated responses and tie them back to the LLM-judge safety verdicts.

→ Blocked (safety-filtered), truncated (length), finish-reason breakdown: See references/guardrails.md

Evaluation Quality

Evaluation results are captured as bizevents (not spans) with event.type == "gen_ai.evaluation.result". Each evaluator emits one bizevent per response, carrying the score, pass/fail label, explanation, and the exact Q&A pair. Use evaluation queries to monitor quality dimensions and surface failed responses with the LLM judge's reasoning. Each bizevent also carries the trace.id of the run that produced the evaluated response, so you can pivot from a quality failure to the spans that caused it.

→ Quality scores, failed evaluations, fail rates: See references/evaluations.md

Empty-State Check

When any signal query returns no rows, do not report "no data found" — first confirm whether the application sends GenAI telemetry at all. These two presence checks show which signal families are present:

If has_genai is zero, report that the application appears not to be instrumented for AI Observability yet — not "no data found". If has_genai is non-zero but a specific family (has_tokens, has_agents, has_tools, has_conversation, has_finish_reason, has_cached_tokens, evals) is zero, only that signal type is missing — for example has_conversation == 0 means session-level analytics are unavailable because the app does not propagate a conversation id, and has_cached_tokens == 0 means prompt-caching telemetry is not being reported. These optional families may use different attribute names depending on the provider/SDK; verify before reporting them absent.

Agent Instructions

Act First, Refine Later

When a user asks for analysis, proceed immediately with sensible defaults. Do not ask for parameter values you can reasonably assume.

Default values when not specified:

Exception — cost prices. Per-model prices are the one input you cannot default (there is no cost field in the data). Ask the user for them before estimating USD; never use prices from memory. See cost-and-tokens.md.

Empty-State Rule

When any signal query returns no rows, run the two presence checks in the Empty-State Check capability above before responding — never reply "no data found". If has_genai is zero, report that the application appears not to be instrumented for AI Observability yet; if only a specific family is zero, say which signal type is missing.

Scope Boundary

This skill covers AI Observability signals for customer GenAI applications only. Product documentation and configuration how-to questions (e.g., "How do I configure the Dynatrace OTLP endpoint?") go to ask-dynatrace-docs — this skill does not contain product configuration how-tos.

Understanding User Intent

Map user requests and prompt-starter phrasings to capabilities:

Common Workflows

Workflow: Cost Investigation

Workflow: Token-Spike / Runaway Investigation

Workflow: Failing Agent Run

Workflow: Guardrail & Safety Review

Workflow: Evaluation Review

References

• references/golden-signals.md — traffic, errors, latency, saturation
• references/llm-signals.md — slowest models, provider latency, token usage by model
• references/cost-and-tokens.md — token usage, spikes, cost-estimation template, usage attribution, prompt-caching economics
• references/agent-signals.md — tool usage, failing agents, step latency, loop/runaway detection, Smartscape agent topology (preview)
• references/conversation-analytics.md — session-level cost, depth, and error rate (gen_ai.conversation.id)
• references/guardrails.md — blocked (safety-filtered) and truncated (length) responses via finish reasons
• references/evaluations.md — quality scores, failed evals, fail-rate, trace correlation (bizevents)