Skip to content

Latest commit

 

History

History
97 lines (69 loc) · 4.02 KB

README.md

File metadata and controls

97 lines (69 loc) · 4.02 KB
Raw


titan

Table of Contents

Introduction

Titan is a sandbox for deploying Prometheus, Thanos, and Grafana to K8s with ArgoCD.

Prerequisites

Titan assumes you're running on a Mac, and have the following installed. If you don't have them, you can install them with Homebrew:

Service Installation
Docker brew install docker
Kind brew install kind
Kubectl brew install kubectl
Helm brew install helm
ArgoCD brew install argocd

Getting started

  1. Create a fork of this repo, and clone it to your local machine.
  2. Open a terminal on the cloned directory and run ./titan init.
  3. When you're done, run ./titan kill to tear down the clusters.

Titan will create local Kind clusters for management and delivery, and deploy ArgoCD to the management cluster. It will also deploy Prometheus, Thanos, and Grafana to the delivery cluster.

Delivery cluster

ArgoCD needs access to the delivery cluster, so it can deploy the applications. To do this, titan uses a kind config file to create a cluster with an apiServerAddress set to the host machine's IP address.

As an added convenience, the build script will attempt to detect your IP address, using the en0 and en1 interfaces, and update the config for you. If it can't detect your IP address, you can set it manually in the delivery/cluster.yaml file.

Repository access

ArgoCD needs access to the repo. GitOps works by pulling the manifests from a git repository, and applying them to the cluster. If you're happy with the manifests in this repo, you can leave them as-is. If you want to make changes, you'll need to fork the repo, and edit the manifests to point to your fork (use https, not ssh):

File Field
management/manifests/application.yaml repoURL
management/manifests/repository.yaml url
delivery/manifests/shared.yaml repoURL

Accessing the UIs

Once the clusters are up, you can access the UIs. You'll need to forward the ports for each service. First, set the context for the cluster you want to access:

$ kubectl config use-context kind-{delivery|management}

Then, forward the ports (in separate terminals):

$ kubectl port-forward svc/{service} -n {namespace} {ports}

List of values to pass into kubectl port-forward:

Service Namespace Ports
argocd-server argocd 8080:443
grafana monitoring 3000:3000
prometheus monitoring 9090:9090
thanos-query monitoring 9090:9090

Custom resources

Helm doesn't support upgrading CRDs in the default /crds directory, so titan uses braid to manage the CRDs for ArgoCD and Prometheus externally. To update the CRDs, run the following:

$ braid update # and commit the changes
$ kubectl config use-context kind-{management|delivery}
$ kubectl apply -f {management|delivery}/shared/crds

Using the CLI

Titan provides a cli to help you raise and lower the stack. It's a simple wrapper around kind, kubectl, and helm, so you can use those tools directly if you prefer.

Command Description
./titan init Create clusters and deploy apps.
./titan kill Tear down clusters.
./titan help Print help text.