Continuous Delivery Foundation Ortelius Project Meetings, 20 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: From Dev to Prod in one step

Description

Presented by Sacha Wharton at the May 20, 2022 Ortelius Visionaries Summit

Automation is king. Learn from Sacha how he has been delivering a streamlined process for moving updates across the pipeline as fast as possible. We've always wanted one 'big button' see how close Sacha has gotten.

A

B

A

And welcome to the autelius visionaries conference, my name is sasha wartin and I'll be presenting dev to prod in one step, I'm a senior devops engineer for cubevisor, which is based in munich, germany and a autelius board member.

A

Just introduced a little bit about tortillas, it's an open source project under the cd foundation and you can think of ortillius as the grand overlord of microservices recording their every move.

A

Just to set up the scene a little bit we're assuming here that we have our source control provider chosen in this. In this case, I've chosen github and we've chosen our cloud provider in this case aws, the foundations of the house, are set up and ready for tenants to move in.

A

We have our environments or stages set up devtest and prod. We have all our containers built we're using eks for all our kubernetes clusters, using captain as our build tool. Tilia says the overlord august cd for our application, parameterization and automation, helm, charts for the case templating and parameterization of the component services, and we are using a springboot java web framework.

A

Here we have a monorepo consisting of some component services within this monorepo, and what are we going to do? Is we're going to break out each one of those services into polyrepos and we're going to copy their code into each of those polyreapers for the respective component service?

A

We will push any parent dependencies to a nexus server, and this will allow any other component services to fetch a dependency or a parent dependency if it needs so.

A

Here we have an example of a polyrepo with the code copied in and obviously all the other component services will be created exactly the same way with their code, and then we can then build each component individually and manage each repo individually.

A

Each codebase repo will be married to a helm chart so that we can parameterize our application. um Help. Charts generally have the final say, and this allows us to really get our component service in in shape before it goes into our environment.

A

Here you can see the three different values files which allows us to move between the different stages where our values div and our values test and our values prod and our values.yaml file, which is for all default configuration. We have a templates folder and in there we usually put our config maps.

A

Here's an example of an environment variable convict map and, just as it says, it's all the environment variables for this particular component service. Here, for example, we have got a config for rabbitmq.

A

It's also using helm templating, which will allow us to replicate this configuration throughout the different stages.

A

Here we can show the values files side by side, so we have our dev values file our test values file in our prod, and you can see. We just replicate our configuration between those different environments.

A

You can see the we have a vault configuration which is using hashicorp vault to store our secrets and in this case it's for our mysql database.

A

You can see there's the spring boot profiles that are active, there's the config server and the branch, so they can fetch config for the component service, any additional com, any additional configuration for the component service and there's our amazon mq configuration.

A

Here we have our values.yaml file, which is serves as a like, I said previously, for all our default configuration for all of across all environments and stages.

A

um For example, I've shown you how you can override your course configuration uh just by adding those configs in your istio configuration you can override any. You can override the core configuration that might have been set in the um component itself by the developers.

A

Then, once we have everything in order with our home charts, we should start seeing a pod being deployed in argo, cd.

A

And you'll see these, you should see some healthy, green hearts like this to show that application is healthy and that each part of the configuration is healthy. If you drill into those blocks, you'll be able to see the community's configuration for, for that particular part of the component.

A

If any of you have any questions et is ready to field anything that you might have on your mind,.

A

Et is officially part of the cd foundation too,.

A

Thank you very much, and, and thank you for the privilege of being able to bring this talk to the hoteliers conference bye for now.