| | | 🔗 Stories, Tutorials & Articles | | | | | | | CERN: Efficient Access to Shared GPU Resources ✅ | | | | | In this article, the authors explain the mechanisms, motivations, and use cases for GPU concurrency on Kubernetes. They explain how GPUs are being used in different areas such as in High Energy Physics (HEP) analysis and Machine Learning (ML) and how the demand for these resources is growing.
They also explain the challenges that come with using these resources such as ensuring optimal utilization, handling spiky workloads, and enabling resource sharing.
The authors also provide an overview of different GPU concurrency mechanisms such as co-locating workloads, time slicing, and GPU Virtualization and explain the advantages and disadvantages of each mechanism.
The article is targeted at organizations that are looking to access and use their data more efficiently and cost-effectively. |
| | | | | | | | | | | Kinsing Crypto Malware Hits Kubernetes Clusters via Misconfigured PostgreSQL | | | | | Kinsing, a known cryptojacking group, has been found exploiting PostgreSQL servers and vulnerable images to gain access to Kubernetes environments.
Microsoft Defender for Cloud researcher Sunders Bruskin found that the group was able to exploit trust authentication misconfigurations on PostgreSQL servers to connect without authentication and execute code.
Additionally, Kinsing targeted servers with vulnerable versions of PHPUnit, Liferay, WebLogic, and Wordpress that were susceptible to remote code execution. The group also scanned for open WebLogic ports to launch malware. |
| | | | | | | | | | Recycling Kubernetes Nodes ✅ | | | | | In this article, the author describes the problems faced while administrating Kubernetes clusters at Yelp, specifically in relation to the replacement of nodes.
The author explains that as the clusters grew in size and complexity, it became increasingly difficult to manually manage the lifecycle of the nodes. The challenges faced included managing the replacement of nodes for security patches and OS upgrades, as well as how to do so without disturbing tenants. The author then goes on to explain the solutions implemented, which included protecting workloads from disruptions using Pod Disruption Budgets (PDBs) and automating node replacement using Clusterman, an open source autoscaling engine.
The goal was to minimize disruption to tenants and reduce the risk of human error in the replacement process. |
| | | | | | | | | | Canary deployments using Kubernetes Gateway API, Flagger and Google Cloud Deploy | | | | | Canary deployment is a technique for testing changes in a production environment by gradually rolling out the changes to a small subset of users before fully deploying them to the entire user base. This allows for real-world testing of the changes and the ability to quickly roll back the changes in the event of any issues.
In this tutorial, you will learn how to use the Kubernetes Gateway API, Google Cloud Deploy, and Flagger to enable faster releases to production for your applications. You will also learn how to use Google Managed Prometheus to observe your application and how to use a 2 namespace setup with a dev namespace for the development environment and a prod namespace where the K8S deployment is done with a gradual traffic shift using Flagger.
This tutorial requires knowledge of kubectl, gcloud, jq and skaffold. |
| | | | | | | | | | Using QuestDB to collect infrastructure metrics | | | | | In this article, the author Steve Sklar, a Cloud Engineer, explains how QuestDB powers the monitoring system of QuestDB Cloud.
Steve is a big fan of QuestDB for its ability to write queries in SQL against a high-performance time series database, which makes writing complex queries a breeze. He explains how QuestDB can be a great choice for ingesting application and infrastructure metrics, it just requires a little more work on the initial setup than the Kubernetes tooling.
The author also provides a detailed architecture of how the system works, including Prometheus remote write and QuestDB ILP and Telegraf. He also provides sample configuration of the yaml file and helm chart, and shows how to query the data with SQL. |
| | | | | |
|
|
