Splunk Observability AI Agent Monitoring Setup

Prerequisites

Tool or access	Purpose	Verify
Bash and Python 3	Run bundled setup and validation helpers	bash --version && python3 --version
Required product/platform access	Inspect or configure the selected target	Complete the documented preflight
Credential files for live modes	Keep secrets out of chat	Verify paths only

Workflow Overview

┌───────────┐   ┌───────────────┐   ┌───────────────┐   ┌─────────────────┐
│ Preflight │ → │ Render/review │ → │ Apply/handoff │ → │ Validate evidence │
└───────────┘   └───────────────┘   └───────────────┘   └─────────────────┘

When to Activate

• Setting up or auditing Splunk AI Agent Monitoring, GenAI telemetry packages, AI agent evaluation telemetry, or adjacent AI infrastructure observability.
• Preview and review the splunk observability ai agent monitoring setup workflow before any live apply phase.
• Diagnose failed prerequisites, generated assets, configuration, or validation evidence.

Scope

Follow the documented read-only or render-first path whenever it is available. This skill does not imply permission to mutate live systems. Require explicit apply flags, protected credentials, and operator review for state changes.

Examples

Inspect the supported setup modes before selecting one:

bash skills/splunk-observability-ai-agent-monitoring-setup/scripts/setup.sh --help

Expected output: usage, supported modes, and required arguments are displayed without changing the target environment.

Inspect validation modes before running completion checks:

bash skills/splunk-observability-ai-agent-monitoring-setup/scripts/validate.sh --help

Expected output: offline, live, and completion options are displayed when the skill supports them; help exits without mutation.

Troubleshooting

Issue	Cause	Resolution
Preflight fails	A required tool or access path is missing	Resolve it before rendering or applying
Rendered assets are incomplete	Required non-secret inputs are absent	Complete intake and render again
Apply is blocked	Review, credentials, or explicit acceptance is missing	Use the documented handoff
Validation is incomplete	Live evidence is unavailable	Record the gap and keep completion open

Overview

Use this skill to build a render-first setup plan for Splunk AI Agent Monitoring and adjacent AI Infrastructure Monitoring. The skill owns orchestration, validation, coverage reporting, and safe delegation; it does not mark UI-only workflows as API-applied.

The renderer always emits a coverage report. Treat coverage-report.json as the source of truth for what is delegated_apply, render, deeplink, handoff, or not_applicable.

Safety Rules

• Never ask for Splunk Observability tokens, Splunk Platform passwords, HEC tokens, LLM API keys, or client secrets in conversation.
• Never pass secrets directly on the command line or as environment-variable prefixes.
• Use file-based secrets only, such as SPLUNK_O11Y_TOKEN_FILE, --o11y-token-file, --platform-hec-token-file, or the delegated skill's file flags.
• Default prompt and response content capture to off.
• Require explicit accept flags before rendering content capture or LLM-as-judge evaluation env vars.
• Do not promise PII redaction, hashing, or truncation. The upstream GenAI utility marks those helpers as not implemented.

Primary Workflow

1. Start from template.example, or create a small JSON/YAML spec with the target realm, collector mode, agent frameworks, and AI infrastructure products. For Kubernetes runtime handoffs, set deployment.workload_kind, deployment.workload_namespace, deployment.workload_name, and deployment.container_name to the exact application workload and container that should receive the environment variables.

2. Render and validate:

   bash skills/splunk-observability-ai-agent-monitoring-setup/scripts/setup.sh \
     --render \
     --validate \
     --spec skills/splunk-observability-ai-agent-monitoring-setup/template.example \
     --output-dir splunk-observability-ai-agent-monitoring-rendered
   

3. Review the generated files:

   • coverage-report.json and coverage-report.md
   • apply-plan.json
   • runtime/python.env and runtime/requirements.txt
   • collector/values-ai-agent-monitoring.yaml
   • handoff.md and doctor-report.md

4. Apply only when explicitly requested:

   bash skills/splunk-observability-ai-agent-monitoring-setup/scripts/setup.sh \
     --apply collector,hec,loc \
     --spec my-ai-agent-monitoring.json \
     --o11y-token-file /tmp/splunk_o11y_token
   

Apply Model

Supported --apply sections are:

• collector delegates to splunk-observability-otel-collector-setup using --apply-k8s or --apply-linux.
• hec delegates to splunk-hec-service-setup --phase apply.
• loc delegates to splunk-observability-cloud-integration-setup --apply log_observer_connect; the Observability UI connection wizard remains deeplink.
• python-runtime and kubernetes-runtime render deterministic application changes and print operator handoffs because the skill does not own the user's application repository or cluster workloads.
• ai-infra-collector delegates to existing product skills when available, such as Cisco AI PODs and NVIDIA GPU, otherwise renders collector overlay handoffs.
• dashboards delegates to splunk-observability-dashboard-builder.
• detectors delegates to splunk-observability-native-ops.

Validation Rules

Validation must fail for:

• send_otlp_histograms: false, because the Agents page requires histogram metrics.
• Prompt/response content capture without --accept-content-capture or accept_content_capture: true.
• Instrumentation-side evaluations without --accept-evaluation-cost or accept_evaluation_cost: true.
• OpenAI evaluations without OTEL_INSTRUMENTATION_GENAI_EVALS_SEPARATE_PROCESS=true.
• Unsupported package names such as splunk-otel-instrumentation-openai-v2 or splunk-otel-instrumentation-vertexai.
• Selected packages whose Python requirement exceeds the declared runtime version.

References

• Read reference.md for the CLI and generated artifact contract.
• Read references/coverage.md when changing coverage status or adding product support.
• Read references/package-catalog.md before changing package names, Python floors, or framework mappings.
• Read references/apply-and-safety.md before changing live apply behavior.
• Read references/troubleshooting.md when updating doctor output.