TLDR DevOps 2025-11-28

If you've ever found yourself debugging across dashboards, logs, and Slack threads just to find an answer, you've felt the limits of traditional monitoring. OpenTelemetry for Dummies helps you shift from chasing symptoms to truly understanding your systems. It provides a clear, vendor-neutral path to move from "what broke?" to that Aha! moment so you can find and fix issues faster than ever. So stop scrambling for answers and see your systems clearly with OpenTelemetry-native observability.Download the free ebook.

Amazon Route 53 now features Accelerated Recovery, offering a 60-minute recovery time objective (RTO) for public DNS record management during service disruptions in the US East (N. Virginia) AWS Region.

The Pulumi Kubernetes Operator (PKO) now integrates with Kargo, a continuous promotion platform, to manage Pulumi stacks as Kubernetes resources with controlled, staged promotions and verification. Kargo updates the Stack resources with new Git references, triggering PKO to reconcile the stack and apply infrastructure changes, enabling systematic control over multi-environment rollouts. A Kargo dashboard provides operational visibility, displaying freight versions, promotion paths, and real-time health information, while custom roles and analysis templates enforce deployment standards automatically.

HashiCorp's Terraform infrastructure-as-code tool was forked to create OpenTofu in response to HashiCorp's license change in August 2023. OpenTofu operates under a Mozilla Public License 2.0 under the Linux Foundation, while Terraform uses a Business Source License restricting some commercial use. Key differences also include OpenTofu's native state file encryption versus Terraform's reliance on external solutions. Platform teams should consider governance requirements, long-term risk tolerance, and alignment with their overall engineering strategy when choosing between the two tools.

Kubernetes depends on external tools like Prometheus, Grafana, and Kube Prometheus Stack to collect and analyze cluster and workload metrics. Strong monitoring focuses on business-aligned signals, consistent labels, correlations with logs and traces, and automated scaling via HPA and VPA.

This post details how to deploy an AI agent on Amazon Bedrock AgentCore with observability powered by OpenTelemetry and Grafana Cloud. OpenLit provides automatic instrumentation for AI frameworks, and Grafana Cloud's AI Observability dashboards can be used to monitor agent performance, debug production issues using distributed tracing, and optimize costs by tracking token usage and model performance. A research assistant agent powered by CrewAI and Meta's Llama 3 is used as an example in the post.

Kyverno 1.16 features a beta version of a new generation of CEL-based policies for Kubernetes and a clear upgrade path. It also introduces namespaced CEL policy types and enhancements in observability, security, and governance. The new Kyverno SDK lays the groundwork for custom tooling and ecosystem integrations.

A new CSI driver leverages rclone to enable Kubernetes pods to mount cloud storage backends as persistent volumes, supporting over 50 storage providers like S3, Google Cloud Storage, and Azure Blob. For enhanced monitoring, the driver includes a Grafana dashboard, and it offers a Remote Control API for programmatic control of mounts within the cluster.

An organization with a multi-cloud approach across AWS, Azure, and GCP reduced fragmented visibility, long outages, and frustrated engineers by adopting OpenTelemetry (OTel), a CNCF graduated project. With OTel, the organization was able to treat observability as part of their platform engineering strategy by unifying a fragmented, multi-cloud ecosystem and giving developers the flexibility to focus on building features instead of wrestling with telemetry.

Fastly has upgraded its DDoS Protection with the Adaptive Threat Engine, improving mitigation accuracy and reducing mean time to mitigation for attacks. The engine now cross-references generated rules against legitimate traffic and uses continuous detection with HTTP Long Polling, enabling attacks to be blocked up to 72% faster while minimizing impact on normal traffic.

Tools like New Relic and OpenTelemetry track application, infrastructure, frontend, backend, and business metrics to optimize performance and user experience for WordPress, as it powers over 810 million sites. WordPress can face performance issues from plugins, servers, databases, or large assets, making monitoring essential.

Amazon CloudFront has added edge and REC metadata, raw query string access, and advanced origin overrides to CloudFront Functions.

EC2 Auto Scaling now supports instance lifecycle policies that let customers retain instances when lifecycle hooks fail or time out, enabling safer and more controlled shutdowns.