Use the skills in the Observe Skills repository to help you onboard your Kubernetes data to Observe.

The following skills are available:

Prerequisites

You must have the Observe CLI installed for these skills to work. To install the CLI, run this command:

curl -fsSL https://raw.githubusercontent.com/observeinc/cli/main/install.sh | bash

To confirm your installation, run:

observe --version

Skill Installation

Run one of the following instructions to get set up. NB: there are some child skills the two top level skills reference.

# Install all necessary skills
npx skills add observeinc/skills -s debug-k8s-collection deploy-k8s-explorer setup-k8s-backend setup-k8s-collection

# Or, install all skills in the skills repo, without prompts
npx skills add observeinc/skills --all

# Or choose skills interactively
npx skills add observeinc/skills

Usage

deploy-k8s-explorer

The example shown here is with Claude Code, but this works for any AI agent. The skill parallels the information collected in the Add Data portal, just collected from your command line/agent.

Information gathering

The skill will first ask a series of questions to understand what you're looking to set up.

You'll get the chance to set up any extra add-ons you desire. For example, you may choose to mask sensitive data.

Observe backend setup

After your configuration has been decided, the skill will set up the observe backend objects needed to collect your data. It will create a token for you to save for future use.

Kubernetes setup

We use a helm chart to configure the Kubernetes Agent. The skill will write a values.yaml file to match how you've configured the agent.

The skill will then prompt you to run some helm instructions to set up collection. Note that the skill is programmed to not run these instructions itself, to prevent unintended infrastructure changes.

Verification

After all resources have been installed, the skill will check that data is flowing, and provide links to view your data.

Handoff

You can continue adding data by instrumenting apps you may be running in Kubernetes or on a host machine. The skill will point you to our Auto-Instrumentation Skill, which will guide you through that journey.

debug-k8s-collection

This skill looks for bugs by walking through a checklist of places where problems may occur, drilling down if an issue is found. After changes, the agent will check if data that is expected to be received is actually being received.

User k8s setup verification

The skill will first check that the context is correct, and the expected pods are running.

Roughly, the skill will check

1. Tenant information and kubectl context
2. Pod statuses
3. Pod Logs
4. Secrets/Tokens
5. Helm chart values
6. Endpoint connectivity issues

Fix verification

The skill will confirm the following:

1. Pods have been re-rolled since config changes were made
2. OTEL collectors are producing data
3. Datasets are receiving data
4. Data in Observe datasets matches data visible with kubectl
5. The data in Observe matches customer expectations

Miscellaneous debugging options

In case the signal is not strong enough from the above investigations, the agent can also prompt the user to turn on Debug Logging and check Self-Monitor Metrics. Finally, there are also some platform specific cases that the skill will check against.

Read more

Use AI to install the Observe Agent on a host

Overview of Data Onboarding and Instrumentation skills