Deploying Mattermost and Kubeflow on Kubernetes with Juju 2.9

Since 2009, Juju has been enabling administrators to seamlessly deploy, integrate and operate complex applications across multiple cloud platforms. Juju has evolved significantly over time, but a testament to its original design is the fact that the approach Juju takes to operating workloads hasn’t fundamentally changed; Juju still provides fine grained control over workloads by placing operators right next to applications on any platform. This is exemplified in our most recent changes to how Charmed Operators behave on Kubernetes.

In recent release candidates of Juju 2.9 (rc7/rc8/rc9/rc10), we’ve done a lot of work to ensure the juju bootstrap process on Kubernetes is as smooth and as universal as possible – meaning it should be easier than ever to bootstrap a Juju controller on a Bring-your-own-Kubernetes!

But don’t take our word for it, deploy yourself some killer apps on a Kubernetes of your choice…

Get Bootstrapped

To get started, you just need:

  • The latest version of Juju from the 2.9/candidate channel
  • Access to an existing Kubernetes cluster (any will do!)
  • A few minutes!

Get started by installing or updating Juju:

# If you're installing from scratch
$ sudo snap install juju --classic --channel=2.9/candidate
# If you're updating an existing Juju install
$ sudo snap refresh juju --channel=2.9/candidate

Now, confirm you have access to a Kubernetes cluster, and bootstrap it! Your KUBECONFIG will be picked up automatically by the bootstrap process, provided you use the same name in the bootstrap command as the name of the kubectl context!

# Check we've got access to a cluster context
$ kubectl config get-contexts
CURRENT   NAME                       CLUSTER              AUTHINFO                              NAMESPACE
          microk8s                   microk8s-cluster     admin
*         super-cool-cluster-admin   super-cool-cluster   clusterAdmin_juju-test_juju-cluster
# Bootstrap the controller on the cluster
$ juju bootstrap super-cool-cluster-admin

Get Charming

When that command returns successfully, you’ll be ready to deploy a range of awesome charms (and charm bundles). Take your pick from Charmhub, or check out some our favourites:

Mattermost

Mattermost is a flexible, open source messaging platform that enables secure team collaboration. Check it out on Charmhub, or if you can’t wait:

# Create a Juju model
$ juju add-model mattermost
# First, deploy PostgresSQL on Kubernetes
$ juju deploy cs:~postgresql-charmers/postgresql-k8s postgresql
# Now deploy Mattermost
$ juju deploy cs:~mattermost-charmers/mattermost --config juju-external-hostname=mattermost.test
# Seamlessly integrate Mattermost and PostgreSQL 🚀
$ juju add-relation mattermost postgresql:db
# Expose the service so you can hit it in your browser
$ juju expose mattermost
# Confirm the deployment was successful:
$ juju status
Model       Controller  Cloud/Region        Version  SLA          Timestamp
mattermost  micro       microk8s/localhost  2.9-rc9  unsupported  13:31:09+01:00
App         Version            Status  Scale  Charm           Store       Channel  Rev  OS          Address        Message
mattermost  mattermost:5.32.1  active      1  mattermost      charmstore  stable    20  kubernetes  10.152.183.54
postgresql  pgcharm:edge       active      1  postgresql-k8s  charmstore  stable    10  kubernetes
Unit           Workload  Agent  Address       Ports     Message
mattermost/0*  active    idle   10.1.215.212  8065/TCP
postgresql/0*  active    idle   10.1.215.211  5432/TCP  Pod configured

Using the example above, you should now be able to get Mattermost on http://10.152.183.54:8065. Your mileage may vary depending on your individual cluster networking setup!

Charmed Kubeflow

If you’re feeling more adventurous, Charmed Kubeflow wraps the 30+ apps that make up Kubeflow with rock-solid ops code. Charmed Kubeflow integrates these charms to provide the best Kubeflow experience, from deployment to day-2 operations.

Extra goodies for Charming Ninjas

We’re also building the foundations of a better future for Juju + Kubernetes. Check out the Future of Charmed Operators on Kubernetes post for more details and try your hand at building a fancy new charm that implements the sidecar pattern! You can check out the developer docs here!

Thank you!

We look forward to hearing your feedback and making Juju even more awesome! You can send us feedback, or get help in a few different ways:

kubeflow logo

Run Kubeflow anywhere, easily

With Charmed Kubeflow, deployment and operations of Kubeflow are easy for any scenario.

Charmed Kubeflow is a collection of Python operators that define integration of the apps inside Kubeflow, like katib or pipelines-ui.

Use Kubeflow on-prem, desktop, edge, public cloud and multi-cloud.

Learn more about Charmed Kubeflow ›

kubeflow logo

What is Kubeflow?

Kubeflow makes deployments of Machine Learning workflows on Kubernetes simple, portable and scalable.

Kubeflow is the machine learning toolkit for Kubernetes. It extends Kubernetes ability to run independent and configurable steps, with machine learning specific frameworks and libraries.

Learn more about Kubeflow ›

kubeflow logo

Install Kubeflow

The Kubeflow project is dedicated to making deployments of machine learning workflows on Kubernetes simple, portable and scalable.

You can install Kubeflow on your workstation, local server or public cloud VM. It is easy to install with MicroK8s on any of these environments and can be scaled to high-availability.

Install Kubeflow ›

Newsletter signup

Select topics you're
interested in

In submitting this form, I confirm that I have read and agree to Canonical's Privacy Notice and Privacy Policy.

Related posts

Easily deploy and manage Mattermost on Kubernetes with the new Juju charmed operator

Introducing the Juju Charms Operator for Mattermost seamless application upgrades and supported features on Linux and Windows. We’re excited to announce our...

What is KFServing?

TL;DR: KFServing is a novel cloud-native multi-framework model serving tool for serverless inference. A bit of history KFServing was born as part of the...

Onboarding and orchestrating network functions with Open Source MANO (OSM)

Do I need to orchestrate my network functions? Well, the answer depends on the price-performance assumptions of your infrastructure and workloads.  It seems...