69 lines
3.2 KiB
Markdown
69 lines
3.2 KiB
Markdown
# Workshop Roadmap
|
||
|
||
## Exercise Map
|
||
|
||
| # | Exercise | Type | Est. Time | Status |
|
||
|---|----------|------|-----------|--------|
|
||
| 01 | Bootstrap ArgoCD | Core | 30 min | ✅ Implemented |
|
||
| 02 | Deploy podinfo via GitOps | Core | 30 min | ✅ Implemented |
|
||
| 03 | MetalLB + Ingress-Nginx (LAN exposure) | Core | 45 min | ✅ Implemented |
|
||
| 04 | Tekton pipeline (image tag bump → GitOps loop) | Core | 45 min | ✅ Implemented |
|
||
| 05 | App upgrade via GitOps | Core | 15 min | ✅ Implemented |
|
||
| 06 | Monitoring: Prometheus + Grafana | Bonus | 60 min | ✅ Implemented |
|
||
|
||
**Total core: ~2.5–3h. Beginners may stop after Exercise 03 (~1h45m).**
|
||
|
||
---
|
||
|
||
## Solution Branches
|
||
|
||
Model: solution branches are **standalone per exercise** (not cumulative).
|
||
|
||
| Branch | State |
|
||
|--------|-------|
|
||
| `solution/01-argocd-bootstrap` | ArgoCD running, root app applied |
|
||
| `solution/02-deploy-podinfo` | podinfo synced via ArgoCD |
|
||
| `solution/03-metallb-ingress` | MetalLB + Ingress-Nginx + podinfo reachable on LAN; CRD `caBundle` drift handling included |
|
||
| `solution/04-tekton-pipeline` | Full Tekton GitOps loop working |
|
||
| `solution/05-app-upgrade` | deployment.yaml bumped to 6.7.0 |
|
||
| `solution/06-monitoring` | Prometheus + Grafana running |
|
||
|
||
---
|
||
|
||
## Verification Status
|
||
|
||
| Exercise | Smoke-tested |
|
||
|----------|-------------|
|
||
| 01 | ✅ Validated (clean VM + bootstrap + root sync) |
|
||
| 02 | ✅ Validated (podinfo app deploy + healthy) |
|
||
| 03 | ✅ Validated (MetalLB + ingress + podinfo URL reachable) |
|
||
| 04 | ✅ Validated after hardening fixes (PSA patch + pipeline runtime fixes) |
|
||
| 05 | ✅ Validated (upgrade/drift workflow over working 04 stack) |
|
||
| 06 | ✅ Validated (Prometheus/Grafana app healthy + Grafana ingress reachable) |
|
||
|
||
Full end-to-end test: completed on `ops-demo-tryout` from clean baseline through 01–06.
|
||
|
||
---
|
||
|
||
## Recent Changes (2026-03-01)
|
||
|
||
- End-to-end smoke test executed in clean tryout environment (`vagrant destroy && vagrant up`).
|
||
- Exercise 04 hardening to make tutorial reproducible:
|
||
- Tekton namespace PodSecurity patch (`pod-security.kubernetes.io/enforce=privileged`)
|
||
- pipeline validate step switched to pure client-side `kubectl create --dry-run=client`
|
||
- clone task now ensures workspace writeability for later task images (`chmod -R a+rwX .`)
|
||
- git clone/push switched to HTTP auth header flow (no URL credential embedding)
|
||
- Exercise 04 docs clarified with explicit PSA semantics and workshop trade-offs.
|
||
- Assignment clarity improvements across docs/01..06:
|
||
- every shell snippet clearly marked as `VM` or `HOST`
|
||
- removed large per-page top callout blocks; context now lives at snippet level
|
||
- Exercise 03 docs expanded with practical explanation around MetalLB manifests and key Kubernetes terms.
|
||
- Exercise 04 docs expanded with:
|
||
- explicit mandatory credential step before PipelineRun
|
||
- clear distinction between Argo wrapper manifest vs full Tekton pipeline manifest
|
||
- Tekton Dashboard + ingress walkthrough
|
||
- `scripts/vm/set-git-credentials.sh` now prints a context-correct PipelineRun path (`/vagrant/...` fallback included).
|
||
- Earlier branch-level fixes remain in place:
|
||
- root recursive discovery
|
||
- MetalLB CRD `caBundle` drift handling
|
||
- Tekton empty `kustomize` drift fix in solution flow
|