Grafana Labs vs OpenTelemetry

Observability Stack: Prometheus, Node Exporter, Grafana

A technical how-to explaining the three-piece observability stack: Prometheus (time-series database that scrapes metrics), Node Exporter (exposes OS-level metrics at a /metrics HTTP endpoint), and Grafana (visualizes Prometheus data as dashboards). The article describes the pull-based model Prometheus uses, the role of Node Exporter as a translator of OS stats, how Grafana queries Prometheus, default ports (Prometheus 9090, Node Exporter 9100, Grafana 3000), basic install commands, a sample prometheus.yml with scrape_interval and job_name, and next steps such as adding scrape targets, writing PromQL queries, and adding Alertmanager for notifications.

• Prometheus is a time-series database and monitoring system that scrapes metrics from HTTP endpoints and exposes its own metrics on port 9090.
• Node Exporter exposes hardware and OS-level metrics at a /metrics endpoint (default port 9100) and is typically installed one-per-machine.
• Grafana queries Prometheus (e.g., via PromQL) to render dashboards and runs by default on port 3000; it stores no metrics itself.

Building an Internal Developer Platform on Azure AKS

This technical article explains how to create an Internal Developer Platform (IDP) using Azure Kubernetes Service (AKS). It outlines core components including AKS as the managed Kubernetes backbone, a service mesh (e.g., Istio or Linkerd) for microservice communication, CI/CD pipelines (Azure DevOps, GitHub Actions, Jenkins) for automated build and deployment, monitoring and logging tools (Azure Monitor, Prometheus, Grafana, Azure Log Analytics), and security/compliance controls (RBAC, Pod Security Policies, Azure Policy, Azure Security Center). The piece describes an example developer workflow from code push to production and summarizes benefits such as increased efficiency, scalability, security, and consistency.

• An Internal Developer Platform (IDP) is a set of tools, processes, and automations that simplifies development, testing, and deployment for developers.
• Azure Kubernetes Service (AKS) is presented as the central managed Kubernetes environment where containerized applications run.
• Service meshes such as Istio or Linkerd are recommended to manage microservice communication, including load balancing, traffic management, and security policies.

Read-Only SRE: Using AI in Production Safely

The author argues for a conservative, observation-first role for AI in production SRE workflows: grant AI read-only access to telemetry (logs, dashboards, events, commits, deployment history, IaC plans) so it can synthesize incident timelines, summarize recent activity, and surface anomalies — but keep production write actions (restarts, scaling, Terraform changes, firewall edits) under human control. The piece frames AI as a fast, always-available “SRE intern” that helps engineers think faster without taking ownership of risky changes. The author acknowledges AI may earn broader operational responsibilities in the future but recommends an onboarding approach that mirrors human engineers: observe, learn, and prove understanding before receiving write permissions. Published on dev.to on 2026-07-10.

• Author recommends giving AI read-only access to production telemetry (logs, events, monitoring dashboards, deployment history, Terraform plans, Git commits) to help with incident troubleshooting.
• Author explicitly advises against allowing AI to perform production writes (restarts, scaling, deleting resources, changing Terraform, updating firewall rules) due to accountability and business-context gaps.
• Author frames AI as a fast 'SRE intern' useful for preparing incident timelines, summarizing dashboards, highlighting anomalies and suggesting possible causes without making changes.

OpenTelemetry Go Logs API and SDK reach release candidate status

OpenTelemetry Go v1.47.0-rc.1 is here. This release promotes the Logs API and SDK to release candidate (RC), the final stage before we provide stable v1 compatibility guarantees. We believe the design is ready, and now we need the community to test …

LLMOps for Compound AI Systems: Observability & Cost

The article argues that most GenAI pilots fail in production due to insufficient system-level engineering rather than poor models. It presents an LLMOps playbook for compound AI systems (embedders, retrievers, vector stores, re-rankers, validators, tool calls, and multiple LLMs) centered on five controls: a model gateway for routing and budgeting, pipeline-level traces for end-to-end observability, semantic caching keyed by query embeddings, lightweight eval gates for safety and quality, and tiered scaling of heavy infrastructure. A concrete engineering example reports a 38% reduction in token spend and 25% lower median latency after implementing a gateway, semantic cache, and tracing. The post includes a short pseudocode example (using qdrant-style vector operations) and an operational checklist for iterating LLMOps as an operating model.

• The article defines five LLMOps controls: model gateway, pipeline-level traces, semantic caching, eval gates, and tiered scaling.
• Author recommends using OpenTelemetry-compatible spans to instrument embed, search, rerank, prompt build, LLM call, and tool call stages.
• A cited engineering example achieved a 38% reduction in token spend and 25% lower median latency after implementing three LLMOps controls.

agent-cost: Measure LLM Usage, Separate Task Attribution

The author describes agent-cost, a small tooling primitive that reads local logs from LLM CLIs (e.g., Claude Code and Codex) to produce auditable, machine-readable usage facts (model, token kind, timestamp, count) and an estimated price. The tool is designed to run with no network calls at runtime, carry a versioned price catalog (with SHA-256 digest), and keep session measurement distinct from task attribution. Unknown or unsupported pricing and ambiguous session-to-task bindings are surfaced (labels like "unpriced" or "lower_bound") rather than silently allocated. The author re-ran the published coding-agent-cost 0.1.0 package and notes a catalog version 2026-07-29 and workflows that validate the measure/v1 protocol and data quality.

• agent-cost reads local logs from LLM CLIs (examples: Claude Code and Codex) and normalizes usage events into facts containing model, token kind, timestamp, and count.
• At runtime agent-cost makes no network calls and declares no Python runtime dependencies; installation from PyPI still requires trust in the supply chain.
• agent-cost carries a versioned pricing catalog with a SHA-256 digest and marks unknown models/prices as 'unpriced' or 'lower_bound' instead of inventing values.