Kubernetes Engine Demos
Where’s Waldo
For the last three months, my GitHub Contributions scoreboard has been looking like an ugly shade of grey.
I had one community managers for Kubernetes reach out and ask if I had been eaten by the all encompasing XYZ. XYZ being her company. Yes I have been working with a world class team with a major client on a fun, and now public, project. Thanks to my partners over at ReactiveOps, I was given the opportunity to work with Google Professional Services on a project that I can only call amazingly awesome.
The team that I work with has been working hard on getting various assets ready for #GoogleNext2018, and some of those assets include Open Source examples using Kubernetes Engine.
Open Source with Google Cloud Platform
As many of you know, Google is a huge proponent and driver of many Open Source projects including Kubernetes. The projects that we released on 07/25/2018 include 15 projects covering various topics focusing on using Kuberenetes Engine on GCP, and connecting Kubernetes Engine on GCP to Kubernetes Engine On-Prem. This release coincided with the news that Google was releasing the product Kubernetes Engine On-Prem.
The Projects
At this point, we have released 15 projects on GitHub that can be all found by clicking here. In those demos and tutorials we covered a range of topics broken into various areas; Databases, Networking, Rolling Updates, Istio, Security, Logging Monitoring and Tracings.
Yes those are a bunch of topics, and there are 15 projects, so this is not a short post.
Databases
GKE Stateful Applications Demo
Leveraging my previous work with Apache Cassandra and Kubernetes, we implemented a new project that demonstrates a Stateful Application, and frankly, the complexities of running a Stateful App on Kubernetes Engine.
Cloud SQL
Running a Cloud SQL Proxy sidecar with a MySQL or Postgres container provides a seamless way to tie an application into GCP, IAM, and a hosted MySQL or Postgres Cloud SQL database.
Logging, Monitoring, and Tracing
Monitoring with Stackdriver
This demo has you setup Monitoring and visualization for a Kubernetes Engine Cluster.
Logging with Stackdriver
You can use this demo to deploy a sample application to GKE and fowards log events to Stackdriver logging. Logs are forwards to a Cloud Storage bucket, and also BigQuery.
Tracing with Stackdriver
Stackdrivers tracing feature is demonstrated by this project. This application deploys a tiny Python Flask application that uses Open Census tracing and integrates into a Pub Sub Topic.
Monitoring with Datadog
Other third party tools can be used to monitor Kubernetes Engine. This code will help you deploy a Datadog Daemonset, and send data to their dashboards. The Datadog agents will be configured to monitor the nginx workload, and ship metrics to your own Datadog account.
Networking
Connect GCP Networks with VPC Peering
This demo uses deployment manager to multiple GKE clusters, and enables VPC Peering between the different clusters. Various applications and service types are deployed to demonstrate different Kubernetes Services.
Connect GCP Networks with Cloud VPN
Instead of using VPC Peering, this project duplicates the above project using a VPN connection. Clusters in one project will be used as stand-ins for “on-premise” clusters, and the VPN will demonstrate remote communication between those clusters, and Kubernetes Engine clusters.
Rolling Updates
This GitHub repository includes three different patterns that can be used to upgrade a GKE cluster. The first is the out of the box upgrade built in; the second two are using patterns that are a great choice for more mission-critical and stateful applications.
In Place Rolling Upgrade
This project demonstrates the out of the box upgrade functionality for Kubernetes Engine.
Expand And Contract Upgrade
Using an “Expand and Contract” pattern, this demo has you upgrade a cluster. The Expand and Contract Upgrade pattern increases headroom by adding 1 or more new nodes to the node pool, prior to starting the upgrade. When the upgrade has completed, the extra nodes are removed.
Blue/Green Upgrade
Using this project you perform a cluster upgrade using a “lift and shift” strategy otherwise know as blue/green. New nodepool(s) are created, and then applications are moved off the old nodes, and on to the new nodepool(s). The old nodepool(s) are then removed.
Security
Network Policies
Network Policies allow fine-grained control and restriction of network communication, internal and external of a Kubernetes Engine Cluster. You can go though this demo and provision a cluster, various pods, and configure their network communication capabilities.
Role Based Access Control (RBAC)
This project allows you to assign different permissions to different users and grant GKE API access to a example application running inside a Kubernetes Engine Cluster.
Securing Containerized Applications
You will deploy multiple instances of the same container image with a variety of security settings to illustrate the use of RBAC, security contexts, and AppArmor policies.
Controlling Levels of Privilege
You will use a combination of RBAC, network policies, security contexts, and AppArmor profiles to enforce appropriate security constraints on three distinct applications.
Istio
Istio with Telemetry
With this project you can deploy Istio, Prometheus, Jaeger and Grafana. This demonstrates how to use all four tools together on a GKE cluser.
Istio with Mesh Expansion
With Istio you can expand the service mesh, as that is a base capability. This demo installs Istio on a GKE cluster, creates a GCE instance with MySQL installed. Then the GCE Instance is added to the mesh.
Istio with Mesh Expansion Across Networks
This demo is the same as above, but uses a VPN connection to demonstrate a mock on-prem connection.
What’s Next
We are still busy working on few projects, but mostly looking forward to working with anyone who would like to help us fix any problems, and improve the communites OSS demos. We love to see contributions and issues!