The article discusses the performance and reliability challenges of the Kubernetes List API when managing large-scale deployments, particularly with over 5,000 nodes and 100,000 pods in a single cluster. It highlights the risks as...
About:
Ahmet Alp Balkan is a software engineer and infrastructure lead with interests in Kubernetes and cloud computing, known for his contributions to open source projects.
Website:
Specializations:
Interests:
Incoming Links:
Outgoing Links:
Subscribe to RSS:
The article explains the different ways Kubernetes terminates workloads, including eviction API, kubelet node-pressure eviction, taint-based eviction, kubelet admission, pod preemption, and node deletion. It also discusses the imp...
0So you wanna write Kubernetes controllers?
2025-01-22 •
The text discusses the process of writing Kubernetes controllers, emphasizing the importance of understanding the API conventions, designing single-responsibility controllers, shaping the Reconcile() method, reporting status and c...
0Notes on OpenAI Kubernetes outage
2024-11-18 •
OpenAI recently experienced a lengthy outage due to a new monitoring agent scraping control plane telemetry, which caused the Kubernetes API servers to become overwhelmed and take down the control plane in most large clusters. The...
The text discusses a low severity incident in Kubernetes clusters at the author's company, involving the node-feature-discovery (NFD) project. The incident occurred after upgrading NFD, leading to bugs, scale issues, and node labe...
0Kubernetes CRD generation pitfalls
2024-09-10 •
The text discusses the pitfalls of generating Kubernetes Custom Resource Definitions (CRDs) using controller-gen. It covers topics such as field validation, zero vs null pitfalls, nested fields validation, field defaulting, and ex...
0Why Kubernetes secrets take so long to update?
2022-12-28 •
The text explains why it takes Kubernetes 60-90 seconds to propagate changes to Secrets and ConfigMaps on the mounted volumes. It delves into the mechanics of these volume types and the configMapAndSecretChangeDetectionStrategy se...
The text discusses the peculiar and undocumented behavior of files on Kubernetes Secret and ConfigMap volumes when it comes to watching changes to these files with the inotify(7) syscall. It explains the challenges of file reloadi...
0Did we market Knative wrong?
2021-06-16 •
The author discusses the mistakes made in the early positioning and messaging of Knative, preventing it from being widely adopted. They believe Knative should have been marketed as a tool for microservices networking on Kubernetes...
The article explains how to code a Go server to serve both gRPC and HTTP/2 endpoints from a single service on Cloud Run. It provides a complete code that multiplexes a sample gRPC service and an HTTP handler over the same port and...
0My review of the System Design book
2021-04-29 •
The book 'System Design Interview – An insider’s guide' by Alex Xu is a good resource for beginners in systems design and distributed systems space. It provides a problem-solving framework and covers topics like rate limiting, con...
0Deploying to Cloud Run with Go
2021-04-29 •
The text is a Go code walkthrough that shows how to deploy and manage services in Cloud Run through its Go API client library. It includes checking if a Cloud Run service exists, deploying a new Cloud Run service, waiting for the ...