Kubernetes schedules GPU workloads by assigning a whole device to a single job exclusively. This one-to-one relationship leads to massive GPU underutilization, especially for interactive jobs, characterized by significant idle periods and infrequent bursts of heavy GPU usage. Current solutions enable GPU sharing by statically assigning a fixed slice of GPU memory to each co-located job. These solutions are not suitable for interactive scenarios since the number of co-located jobs is limited by the size of physical GPU memory. Consequently, users must know the GPU memory demand of their jobs before submitting them for execution, which is impractical.

Understand Kubernetes Events and learn to use kubectl events to monitor and troubleshoot your cluster’s issues effectively.

The nuances of how users and groups are configured in Kubernetes and how the role-based access control (RBAC) mechanism applies for them.

Argo Event is a Kubernetes based event automation engine. It is part of the Argo project. Argo Events can be used with or independent of other projects in Argo.

I will be writing a series of articles on Argo Events; in these articles I will be looking at how we can use Argo Event to automate process within and without a Kubernetes cluster.

For this first article in this series, we will examine Argo Events core concepts, installation and provisioning different event buses which Argo Event uses to forward events to their sink. Finally we will look at setting up a webhook event flow to verify our setup.

Know the differences between ConfigMaps mounted as Volumes and ConfigMaps defined as environment variables.

A database does not need a write-ahead log (WAL) to achieve durability. A database can write its long-term data structure durably to disk before returning to a client. Granted, this is a bad idea! And granted, a WAL is critical for durability by design in most databases. But I think it's helpful to understand WALs by understanding what you could do without them.

Applying Quality of Service techniques at the application level

What is it? Why is it used? And why is it important in the context of optimization and reliability engineering? Bonus: a browser app that lets you play with data.

We have :

- 15 terragrunt.hcl files for every service.
- Around 900 terragrunt.hcl files (15*60) in total.

Updatecli is a command-line tool used to define and apply update strategies. It reads a manifest then works into three stages:
1 Source, which describes where a piece of information, to be used, is retrieved from.
2 Target, which describes what to update using information defined by the source.
3 Condition, which defines a condition, required to be satisfied, to update the target.

Deciding how, when, where to update information is hard.

There are many tools that can apply continuous delivery or continuous deployment. We configure our infrastructure with ansible playbooks, puppet manifest, helm chart, and others. We rely on configuration files to specify the version we need to install. Unfortunately, too often those files are manually updated.

Because it’s hard to automatically detect what information to update, and when.

The logic that manipulates information from a configuration file is defined outside that configuration file. Information comes from different sources like maven, docker, files, git repository, and elsewhere. Before modifying information, we should validate our assumptions.

Updatecli allows combining blocks, aka plugins, to specify what information needs to be updated, when, and where. We can easily implement the workflow that suits our needs.

Atuin replaces your existing shell history with a SQLite database, and records additional context for your commands. With this context, Atuin gives you faster and better search of your shell history.

Additionally, Atuin (optionally) syncs your shell history between all of your machines. Fully end-to-end encrypted, of course.

Since the 1.7 release, the OpenTofu community and core team have been hard at work on much-requested features, making .tf code easier to write, reducing unnecessary boilerplate, improving performance, and more

RFC1178 Choosing a Name for Your Computer