| name | ocp-observability |
| metadata | {"author":"rhoai3-demo","version":"1.0.0","platform-family":"ocp","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"OpenShift Platform"} |
| description | Use when documenting, reviewing, or rebuilding OpenShift Container Platform observability guidance from official OCP documentation: Observability overview, monitoring stack, user-defined project monitoring, metrics, customized alerts, Prometheus, Alertmanager, Thanos Querier, logging, infrastructure/application/audit logs, log collection, log visualization, log forwarding, log storage, Cluster Observability Operator, customizable monitoring stacks, and release-cadence handoffs to separate Red Hat observability documentation. Do NOT use for Red Hat OpenShift AI observability stack configuration, TrustyAI model monitoring, deployed model metrics, or RHOAI dashboard observability; use the relevant rhoai-* skill. Do NOT use for GitOps-managed Grafana Operator, Grafana instances, datasources, dashboards, OAuth routes, or Grafana RBAC; use ocp-grafana-operator. Do NOT invent monitoring, logging, or COO custom resource fields from OCP overview pages alone.
|
OCP Observability
Use this skill to ground OpenShift Container Platform observability guidance in
the official OCP documentation for the active baseline in
docs/PLATFORM_BASELINE.md.
Source Grounding
Read references/source-capture.md before using product behavior. Official Red
Hat documentation is product authority. This skill captures the OCP
observability model, the default monitoring posture, the logging scope, and the
Cluster Observability Operator overview.
The OCP observability sources intentionally hand off detailed monitoring,
logging, and COO behavior to separate Red Hat documentation sets. Before
authoring manifests, verify the relevant separate product documentation or the
active cluster schema.
Demo Observability Posture
For this AWS-hosted RHOAI demo:
- Use OCP platform observability as the base layer before adding RHOAI model,
TrustyAI, or application-specific observability.
- Keep platform metrics, user workload monitoring, alerting rules, logging, and
COO resources GitOps-managed once active implementation artifacts exist.
- Do not claim that a metric, alert, dashboard, log pipeline, trace, or
observability stack exists unless the implementation and validation evidence
exist.
- Keep OpenShift platform observability separate from Red Hat OpenShift AI
observability. Use
rhoai-observability,
rhoai-model-management-monitoring, and rhoai-monitoring-trustyai for
RHOAI-specific surfaces.
- Treat log output configuration, external log sinks, telemetry endpoints, and
credentials as sensitive live-environment concerns.
- Let OLM and the Cluster Observability Operator own generated operand images,
CSV
relatedImages, copied CSVs, and generated observability resources.
Diagnose image/version mismatches through Subscription and CSV inspection,
then use operator lifecycle policy or product baseline alignment instead of
generated image patches.
OCP Observability Model
Use the official docs to frame:
- Red Hat OpenShift Observability: real-time visibility, monitoring, and
analysis across metrics, logs, traces, and events.
- Observability components: Monitoring, Logging, Distributed tracing, Red
Hat build of OpenTelemetry, Network Observability, and Power monitoring.
- Unified solution: observability components connect open-source tools and
technologies to collect, store, deliver, analyze, and visualize data.
- Release cadence boundary: except for monitoring, Red Hat OpenShift
Observability components have release cycles distinct from core OCP releases.
- Monitoring: OCP includes a preconfigured, preinstalled, and self-updating
monitoring stack for core platform components. Cluster administrators can
optionally enable monitoring for user-defined projects.
- Logging: logging collects, visualizes, forwards, and stores log data for
troubleshooting, performance analysis, and security threat detection.
- Cluster Observability Operator: COO is optional and supports creating and
managing highly customizable monitoring stacks with more tailored namespace
views than the default OCP monitoring system.
Workflow
- Confirm the active OCP baseline in
docs/PLATFORM_BASELINE.md.
- Read
references/official-doc-extraction.md.
- Identify whether the task concerns:
- OCP-level observability narrative or component mapping
- default platform monitoring
- user-defined project monitoring
- logging collection, visualization, forwarding, or storage
- Cluster Observability Operator positioning
- detailed monitoring, logging, or COO CR fields that require separate docs
or active schema verification
- RHOAI-specific observability, which belongs in the relevant
rhoai-*
skill
- For manifests, verify all API versions, CRDs, fields, namespaces,
Operators, RBAC, credentials, and output endpoints before committing.
Classify image fields as repo-owned or operator-generated before proposing
any image pin.
- For live operations, use the repo environment guard and pair this skill with
env-troubleshoot, env-manage-resources, or env-deploy-and-evaluate.
- Validate the output with
references/validation-checklist.md.
Related Skills
- Use
rhoai-observability for the Red Hat OpenShift AI observability stack,
dashboard visibility, DSCInitialization.spec.monitoring, and
redhat-ods-monitoring.
- Use
rhoai-model-management-monitoring for deployed model metrics, KServe,
vLLM, NIM, and Grafana model dashboards.
- Use
rhoai-monitoring-trustyai for TrustyAI fairness, bias, drift, and
TrustyAIService workflows.
- Use
project-gitops-authoring for repo-specific GitOps implementation.
- Use
project-manifest-review for structural and security review of
Kubernetes and OpenShift manifests.
- Use
ocp-opentelemetry for Red Hat build of OpenTelemetry Operator,
Collector, instrumentation, telemetry pipelines, and OpenTelemetry metrics
integration.
- Use
ocp-distributed-tracing for Red Hat OpenShift distributed tracing
platform, Tempo Operator, TempoStack, TempoMonolithic, tenants, trace
RBAC, Jaeger UI, and distributed tracing UI plugin behavior.
- Use
ocp-grafana-operator for GitOps-managed Grafana Operator, Grafana
instances, datasources, dashboards, OAuth routes, and Grafana RBAC.
- Create narrower
ocp-* skills for Network Observability or Power monitoring
when official sources are provided and implementation needs them.
References
references/source-capture.md
references/official-doc-extraction.md
references/validation-checklist.md
examples/observability-review-patterns.md