Press "Enter" to skip to content

Category: DevOps

MLOPS in R with GitHub Actions

David Smith explains MLOPS and GitHub actions in a talk:

In the talk, I demonstrate the process in action (the demo starts at the 14:30 mark in the video below). I used Visual Studio Code to edit the app.R file in repository, and then pushed the changes to GitHub. That immediately triggered the action to deploy the updated file via SSH to the Shiny Server, running in a remote VM. Similarly, changes to the data file or to the R script files implementing the logistic regression model would trigger the model to be retrained in the cluster, and re-deploy the endpoint to deliver new predictions from the updated model.

Click through for a quick summary, link to the repo, and embedded video of the talk.

Comments closed

Release Rollback with Helm

Andrew Pruski shows the secret of how Helm lets you roll back releases even when deployments are deleted:

If we rollback with kubectl rollout undo the pods in the newest replicaset are deleted, and pods in an older replicaset are spun back up, rolling back the upgrade.

But there’s a potential problem here. What happens if that old replicaset is deleted?

If that happens, we wouldn’t be able to rollback the upgrade. Well we wouldn’t be able to roll it back with kubectl rollout undo, but what happens if we’re using Helm?

Read on to learn how the whole thing works.

Comments closed

Retrieving Secrets from Azure DevOps Pipelines

Gavin Campbell shows how you can pull secrets out of an Azure DevOps Pipeline:

For secrets created in the Azure DevOps UI, whether pipeline-scoped or in a variable group, it is not so simple to retrieve the variables after creation. This might be required for a number of reasons, most often troubleshooting. The need to do this is often an indicator that the project should have been using an Azure Key Vault in the first place.

Previously it was necessary to jump through some hoops to access secret variables, but it turns out this is no longer required. It also appears the recommended approach of mapping secrets to environment variables is currently not working for secret variables from variable groups.

I second the notion of using Key Vault for secrets management.

Comments closed

Deploying ADF via Azure DevOps

Kamil Nowinski has part two on a series about releasing Azure Data Factory code:

Struggling with #ADF deployment? adf_publish branch doesn’t suit your purposes? Don’t have skills with PowerShell? I have good news for you. There is a new tool in the market. It’s a task for Azure DevOps Release Pipeline to deploy whole ADF from code (JSON files) to ADF instance in Azure. Behind the scenes, it runs the PowerShell module which does all job for you.
Sounds unbelievable? But it’s real! Check it out for yourself.

Click through for the video.

Comments closed

Publishing Azure Data Factory via Azure DevOps

Kamil Nowinski shares how to deploy Azure Data Factory flows via Azure DevOps:

Struggling with #ADF deployment? adf_publish branch doesn’t suit your purposes? Don’t have skills with PowerShell? I have good news for you. There is a new tool in the market. It’s a task for Azure DevOps Release Pipeline to deploy whole ADF from code (JSON files) to ADF instance in Azure. Behind the scenes, it runs the PowerShell module which does all job for you.
Sounds unbelievable? But it’s real! Check it out for yourself.

Click through for a video.

Comments closed

Operations Testing with Pester

Sheldon Hull takes us through using Pester to automate operations tasks:

In my example, let’s start small and say you just have PowerShell, and some servers.

What I’ve discovered is that to actual validate DevOps oriented work is completed, you typically go through the equivalent of what a Cucumber test would have. This “checklist” of validations is often manually performed, lacking consistency and the ability to scale or repeat with minimal effort.

Consider an alternative approach to helping solve this issue, and expanding your ability to automate the tedious testing and validation of changes made.

Read on for an example as well as some additional thoughts from Sheldon.

Comments closed

Applying the Principles of Site Reliability Engineering

Sheldon Hull has an essay on site reliability engineering in practice:

I’ve always been focused on building resilient systems, sometimes to my own detriment velocity wise. Balancing the momentum of delivery features and improving reliability is always a tough issue to tackle. Automation isn’t free. It requires effort and time to do correctly. This investment can help scaling up what a team can handle, but requires slower velocity initially to do it right.

How do you balance automating and coding solutions to manual fixes, when you often can’t know the future changes in priority?

This is personal experience rather than prescriptive guidance. Very interesting personal experience.

Comments closed

CI/CD with Databricks

Sumit Mehrotra takes us through the continuous integration story around Databricks:

Development environment – Now that you have delivered a fully configured data environment to the product (or services) team in your organization, the data scientists have started working on it. They are using the data science notebook interface that they are familiar with to do exploratory analysis. The data engineers have also started working in the environment and they like working in the context of their IDEs. They would prefer a  connection between their favorite IDE and the data environment that allows them to use the familiar interface of their IDE to code and, at the same time, use the power of the data environment to run through unit tests, all in context of their IDE.

Any disciplined engineering team would take their code from the developer’s desktop to production, running through various quality gates and feedback loops. As a start, the team needs to connect their data environment to their code repository on a service like git so that the code base is properly versioned and the team can work collaboratively on the codebase.

This is more of a conceptual post than a direct how-to guide, but it does a good job of getting you on the right path.

Comments closed

Reading Azure DevOps Results in Powershell

Mark Broadbent doesn’t let the lack of an official Powershell module get in the way:

In my post Using Azure CLI to query Azure DevOps I explained how you can use the Azure CLI to query Azure DevOps so you can obtain useful information on builds, releases, and other useful information. The solution required a certain level of skill with JMESPath to manipulate your result sets -which as explained can be a little confusing.

However once you have a bare bones result set, it is likely that you will want to consume these results in a more user-friendly environment such as PowerShell so that you can build upon these data sets. I thought this would be an easy thing to do, but as you will see below it was anything but.

Read on for some thoughts and a sample script.

Comments closed

Scripting and Deploying SQL Agent Jobs

Alex Yates shows how you can incorporate SQL Agent jobs in your CI/CD process:

Basically, we need to put all the SQL Agent Job .sql scripts into a git repo. Then we need a PowerShell script that executes each .sql script against the necessary target databases. If you use SSDT, you might prefer to use a post deployment script to do this. That bit should be reasonably straight forward. I’ll leave that as a task for the user since I’m short on time.

You probably want to put some thought into whether your agent jobs are scoped to a particular database, general server admin for a specific server, or whether you want them to be standardised across many servers since this may affect where you choose to put your jobs ion source control and on what schedule you want to deploy them.

It may also make sense to set up MSX if you have a central server. That would make Agent job deployment easier and you can still script out which sets of servers get which jobs.

Comments closed