Objectives & prerequisites

What Hello Pico on Manifold teaches and what you need before starting.

Learning objectives

NoteLearning objectives
  • Deploy a small Manifold RuntimeEnvironment onto a local Kubernetes cluster.
  • Host a Pico engine Pod inside that RuntimeEnvironment as the in-process host for one Pico Element.
  • Declare a single Wrangler Channel and subscription that wires one event into the hosted Pico, using the topology YAML fragment shipped with this lab.
  • Send exactly one event onto that Channel with a small Control Surface command (a kubectl exec invoking wget inside the cluster).
  • Capture the single log line the Pico prints in response as the lab’s runtime Observation, and store it alongside the applied manifests as hello-world-pico-on-manifold/observation.txt.

Prerequisites

NotePrerequisites
  • Completion of the Hello Pico lab — the greeting value in this lab reuses the same Hello, Pico! string.
  • Familiarity with the Manifold course vocabulary (RuntimeSubstrate, RuntimeEnvironment, Channel, EventType, Observation).
  • A local Kubernetes cluster; see the Kubernetes setup foundation for the canonical minikube install, PATH checks, and first cluster bring-up. helm is not required for this lab.

Environment

The canonical local Kubernetes setup — supported OS, install commands for docker, minikube, and kubectl, PATH checks, and a first cluster bring-up — lives in the Kubernetes course setup foundation. Complete that page first so docker, minikube, and kubectl resolve on your PATH.

Lab-specific additions on top of that canonical setup:

  • Reference minikube pin: v1.38.1 was used to author this lab (any v1.34+ should work).

  • Bash utilities: grep, cat, mkdir, cp — invoked by the walkthrough and by verify.sh.

  • Estimated time: 45–60 minutes (of which the cluster start accounts for roughly 2–5 minutes).

  • Cleanup step (also repeated at the end of the walkthrough):

    minikube delete --profile manifold-lab
rm -rf build/hello-world-pico-on-manifold work/hello-pico-on-manifold

Pinned versions (reference)

The walkthrough was authored and verified against these versions. Newer versions are expected to work; older Kubernetes versions may lack API resources used by the shipped manifests.

  • Kubernetes API version in minikube: v1.31.0
  • Container image used as the Pico engine stand-in: busybox:1.36 (chosen for its small footprint and the nc / wget utilities the lab uses to send and receive one event over the Channel).
NoteWhy busybox for the Pico engine

The Pico engine in this lab is deliberately implemented as a small busybox-based Pod running a single-shot listener. It plays the role of a Pico engine hosted by a Manifold RuntimeEnvironment: it consumes one event from a Channel and produces one Observation back onto its own log. It is not a full Pico engine implementation; Phase 7 of the constructive-realization contract explicitly says no new runtime implementation is introduced here. Later curriculum waves may swap the stand-in for a real Pico engine image without changing the surrounding topology.

One-time setup (optional)

The walkthrough uses a dedicated minikube profile named manifold-lab so it does not interfere with other clusters (for example the crossplane-lab profile from the Hello Pico on Kubernetes lab). If you prefer to reuse an existing minikube cluster, drop --profile manifold-lab from every command. If you use a different Kubernetes distribution (kind, k3d, docker-desktop), all kubectl steps below still apply.