Nathan Nellans

One of my beefs with terraform is it’s trying so hard to be flat. Some of us want very dynamic scripts. But when you create a k8s cluster, form the kubexonfig file, then need to use the kubexonfig file to load in the rest of the parts like ingress, having a depends_on would be nice for providers. I did cheat a little and abuse the variable system to force the provider to wait, but then you run into what about a workspace that will have kubernetes set to false and get a zero count for your k8s module? Then it gets angry because the module output isn’t there. I still like Chef. Chef let me do what I want when I want. Just was bad on creating the underlying infrastructure but great at configuring it once it was up. *shrug* #terraform #kubernetes #k8s #eks #automation #burgerkingmodel

3

1 Comment

Another quality of life pre:invent release - this looks like it will streamline complex state machines. https://lnkd.in/enc2Qsgi

1

Imagine you have the following requirement: some pods need to be placed on specific nodes dedicated to compute intensive, memory intensive, and storage intensive workloads. How do you ensure these pods are scheduled correctly and not placed on other nodes? At the same time, how do you ensure that no other pods are placed on these optimized nodes? My latest article, “Kubernetes Pod Scheduling Explained: Taints, Tolerations, and Node Affinity”, tackles this exact scenario. https://lnkd.in/d3X7w2Bi #kubernetes #devops #cloud

4

The Kubernetes AI toolchain operator (KAITO) is an open-source project also available as an AKS add-on. The KAITO add-on for AKS enables running specialized machine learning workloads like LLMs on more cost-effectively and with less manual configuration.

2

terraform plan -light Add a terraform plan -light flag such that only resources modified in code are targeted for planning. https://lnkd.in/gGU5UJ94

4

Unit 42 researchers test container escape methods and possible impacts within a Kubernetes cluster using a containerd container runtime. https://lnkd.in/g4pwwxen

4

GitOps simply allows faster deployments by having git repositories in the center offering a clear audit trail via git commits and no direct environment access. ✍️✍️✍️ Deploying with GitOps https://lnkd.in/gr5kbwhN

6

The most promising thing I’ve seen from Microsoft Build is AKS Automatic. Is this finally kubernetes for the masses without the burden and overhead of managing k8s clusters? Can’t wait to try it #cloudnative #k8s #kubernetes https://lnkd.in/eBCgVsTb

14

🚀 Exciting news for DBAs! MySQL 8.4 has officially dropped as a Long-Term Support (LTS) release, promising enhanced stability and a steady stream of bug fixes until early 2026. From improved security features to the much-anticipated Clone Plugin updates, this release is set to be a game-changer. Curious about all the updates? Dive into the details on the Release Notes. Stay tuned for more insights as I explore these features in upcoming posts! #MySQL #DatabaseManagement #TechUpdates 🛠️✨ https://lnkd.in/egm9Hevv

7

Some early next-day thoughts about #DASH2024 in New York City this year: 👌 Logs Workspaces look super promising ☀️ Flex Logs go GA (can't wait for Flex Traces 👀) 💥 On-call (👋 Opsgenie & PagerDuty) 🙈 GitHub is on Datadog (and OpenTelemetry!) 💸 Paypal is on Datadog too! 💜 OpenTelemetry + Datadog are getting closer than ever 🎉 Serverless improvements 🥂 Product Analytics revamp, becoming a proper competitor with products like Amplitude 🤖 Bits AI is being used in more and more products 👏 App Builder (custom UI components you can build yourself into the Datadog UI) is better than ever. 🤯 Live Debugger - merges the concepts of debugging and tracing 🔐 Datadog secures the entire software development cycle My biggest takeaway though is that Datadog is bigger than ever. But they are still developing products at an incredible pace and quality. From a Product perspective, they are hitting the nail on the head over and over again. Datadog releases things as early as possible as "Beta", "Private Beta" etc. This may frustrate users (as early products are immature), but it actually provides incredible opportunities for feedback. A bit of an observation: There are a lot more instances where Datadog goes past this "understanding platform" idea. It's now moving to be able to take action. This can be a slippery slope but I do think in many cases it makes loads of sense. A roundup from Datadog themselves of the keynote:

25

4 Comments

Learn how to use Kyverno, the native policy engine for Kubernetes, to enforce policies, validate configurations, and manage resources dynamically within your Kubernetes clusters. https://lnkd.in/gaDZkFm9

48

2 Comments

Terraform's Import block functionality will help give a lot of companies the ability to bring their existing infrastructure under a unified code base. Now with the ability to use for_each inside their import block this video shows a cool way to utilize that feature to build on top of your existing pipelines without having to re-engineer anything.

4

Here are my thoughts on #observability, in no particular order: If your application doesn't produce any #logs (#metrics, #spans, you name it), troubleshooting becomes impossible. You may not even know if it's running. This is especially problematic when your application logs only during usage, but it's not being used. On the flip side, if your application generates too many logs, it can also be challenging to troubleshoot, even if you have log querying tools. Logging "Health check succeeded!" every second for routine health checks contributes to log overload, compounding the previous issue. When log entries are excessively long (structured or not, someone will have to look at your logs at some point). When structured logs contain large, embedded JSON objects. When structured logs embed lengthy lists with line breaks (which are encoded). When stack traces dive into the internals of your HTTP server (or any other third-party library). When the same information is logged repeatedly across 15 layers of error handling. When you cannot attribute an error to an invocation or input parameters. When you log sensitive information. Happy logging!

8

🛡️ When the Going Gets Tough, the Tough Build k8sPulse! I recently found myself needing an emergency Kubernetes monitoring tool—one that could run with just kubectl access, especially in those moments when everything else seems to fail. This need gave me a crazy idea: why not create my own emergency Kubernetes dashboard? What started as a simple Bash script grew into something much more! 🤖 I transformed my emergency dashboard idea into a fully functional, easy-to-use CLI called k8sPulse. It’s built in Python, leveraging the native Kubernetes client and enhanced with Pandas and Matplotlib for beautiful insights. I even challenged myself with design — despite not having any background in UX/UI or color theory, I decided to design the entire product, including its modern homepage: https://lnkd.in/gaap6b2e. 🛠 Want to try it out? If you have your kubeconfig ready, just install it with pipx, then, in any directory: k8spulse This will generate a full report in a docs folder. And if you have an OpenAI API key, you can enable intelligent recommendations! 💡 Check it out, and I’d love to hear your thoughts! #k8sPulse #OpenSource #Kubernetes #Python #CLI #AI #EmergencyMonitoring #SelfTaughtDesign

12

Most cloud architectures are decent. easy copy AWS and Azure free best practice. Most execution is below-average. Easy to read how, often hard to do. Focus on the below-average part.

3

Announcing support for MariaDB Enterprise Server on IBM Power We’re thrilled to announce that IBM® and MariaDB have joined forces to introduce support for MariaDB Enterprise Server on IBM Power® architecture (ppc64le). Clients leveraging MariaDB Enterprise Server on IBM Power benefit from enterprise reliability, stability, long-term support, and greater operational efficiency when managing large database deployments for business- and mission-critical applications. https://lnkd.in/d24CcYxF

3

📢 Dynamic Scaling is now GA in Azure CosmosDB🎉 1️⃣ Save up to 70% on autoscale costs for non-uniform workloads 2️⃣ Independent scaling per region and partition 3️⃣ Achieve 99.999% high availability 🤯 Learn more👉https://lnkd.in/eDTWxj_y Well done Jay Gordon and team🎉 #data #database #programming #development #bigdata #ai #microsoft

24

Welcome to the home of the Azure Proactive Resiliency Library v2 (APRL). The purpose of this site is to provide a curated catalog of resiliency recommendations for workloads running in Azure. Many of the recommendations contain supporting Azure Resource Graph (ARG) queries to help identify non-compliant resources. #Azure #Microsoft #Reliability

5

New ways to manage SPO version history and a much-requested site version storage report are here #sharepoint

4