CloudOctober 202612 min read
State of FinOps 2026: What 98% Managing AI Spend Means
The State of FinOps 2026 survey of 1,192 practitioners says 98% manage AI spend. What the figures show on AI cost visibility, ROI, FOCUS and token metrics.
Read MoreBlog
Cloud
Articles comparing cloud providers and managed services — practical tradeoffs for running production workloads.
29 articles
CloudOctober 202611 min read
OpenCost Inference Cost per Million Tokens for Self-Hosted vLLM
OpenCost 1.121.0 read closely: allocation versus usage cost per million tokens, the label join that zeroes it, PromQL for utilisation and the API break-even.
Read More
CloudOctober 202613 min read
AWS Lambda MicroVMs for AI Agent Sandboxes: Hooks and Limits
AWS Lambda MicroVMs run one Firecracker VM per agent session, with suspend and resume. Hooks, quotas, a Fargate comparison and a Monday test plan.
Read More
CloudOctober 202612 min read
Gemini Cloud Assist Investigations: Proactive RCA and FinOps
Gemini Cloud Assist's alert-triggered RCA and 24/7 cost-anomaly agent, read from Google's docs: April 2026 gating, agent identity, Eventarc and failure modes.
Read More
CloudOctober 202611 min read
Amazon EKS MCP Server: SigV4, CloudTrail and Write-Mode Risks
The managed Amazon EKS MCP server read closely: the SigV4 proxy, the three IAM permissions, what CloudTrail records, and why production should stay read-only.
Read More
CloudOctober 202612 min read
AWS MCP Server in Claude Code: OAuth, SigV4 and IAM Scoping
How to connect Claude Code to the managed AWS MCP Server with OAuth or the MCP Proxy for AWS, and scope IAM so an agent cannot delete what it can read.
Read More
CloudOctober 202612 min read
Azure SRE Agent: Tool Permissions, Hooks and Approval Gates
How Azure SRE Agent stacks hooks, allow/ask/deny tool policies, connector approvals and run modes, where Ask quietly stops asking, and what to configure first.
Read More
CloudOctober 202612 min read
AWS DevOps Agent: Incident Investigation, Cost and Accuracy
A close reading of AWS DevOps Agent: how it triages and investigates alerts, what per-second billing adds up to, and how to test its 94% accuracy claim.
Read More
CloudOctober 202613 min read
High Availability Design: Redundancy, Failover and the Nines
How to design for high availability: what each nine allows in minutes, series versus parallel availability maths, failover detection time and a keepalived and nginx sample.
Read More
CloudOctober 202611 min read
Service Mesh vs API Gateway: Differences and When to Use Each
Service mesh vs API gateway: north-south versus east-west traffic, sidecar versus ambient mesh, real nginx and Istio config, cost, and when you need neither.
Read More
CloudOctober 202613 min read
Design a Video Streaming System: HLS, CDN and Adaptive Bitrate
How video streaming works and how to design it: a transcode ladder, HLS segments and manifests, object storage, a CDN, adaptive bitrate logic, signed URLs and live versus VOD, with the storage and egress arithmetic shown.
Read More
CloudOctober 20269 min read
How a CDN Works: What to Cache at the Edge and What Not To
A concept explainer on points of presence, anycast routing, cache keys, Cache-Control headers, purging, and the responses that must never reach a shared cache.
Read More
CloudOctober 20269 min read
Horizontal vs Vertical Scaling: Which to Pick First
Scale up first, scale out when you need redundancy. How statelessness, session handling, the database and cost shape decide between horizontal and vertical scaling.
Read More
CloudOctober 202614 min read
Cloudflare Containers: 6x Faster Starts for Agent Sandboxes
Cloudflare rebuilt Containers for agent sandboxes: a 648 ms median start, runtime image and instance selection, filesystem snapshots, and a 31 December 2026 migration deadline.
Read More
CloudSeptember 202611 min read
Indonesia Sovereign AI Data Centres: What Actually Changes
Sovereign AI capacity is landing in Indonesia. Two things genuinely change for a team shipping this quarter, three widely assumed things do not, and a contracted megawatt is not an instance you can boot.
Read More
CloudSeptember 202611 min read
Jakarta GPU Capacity: VRAM Sizing for Indonesian AI Teams
Sizing a GPU for an Indonesian AI product is arithmetic: weights, KV cache and the overhead nobody budgets for. What you can rent this quarter, and why quantisation moves the answer.
Read More
CloudSeptember 202612 min read
PaaS vs VPS in Indonesia: An Honest IDR Cost Breakdown
A managed Indonesian PaaS against a Jakarta VPS at matched 2 vCPU and 2 GB in real rupiah, plus the column price comparisons omit: patching, TLS, tested backups, monitoring and on-call, counted in hours.
Read More
CloudSeptember 202612 min read
Kubernetes HPA Autoscaling: Surviving a PaaS Traffic Spike
Autoscaling was on, and the opening seconds of the spike were still served by the pods I already had. The real Kubernetes HPA timeline on a managed PaaS, default by default.
Read More
CloudSeptember 202612 min read
Jakarta vs Singapore PaaS Latency for Indonesian Users
Why hosting in Singapore costs Indonesian users milliseconds: the fibre floor you can compute, the four round trips in one cold HTTPS request, and how to measure both.
Read More
CloudSeptember 202612 min read
Helipod Review: Deploying Next.js on an Indonesian PaaS
A documentation-and-arithmetic review of Helipod: how Helipack builds a Next.js repository with no Dockerfile, what Rupiah and QRIS billing really cost, and the workloads I would still keep on a VPS.
Read More
CloudSeptember 202613 min read
Helipod vs Dokploy: Managed PaaS or Self-Hosted VPS
Helipod is a managed Indonesian PaaS with rupiah billing; Dokploy is a PaaS you host yourself. A comparison of cost, licence terms and who gets paged.
Read More
CloudAugust 202613 min read
Claude Code on Bedrock, Google Cloud and Microsoft Foundry
Running Claude Code on a cloud provider keeps the CLI intact but drops a specific list of features. Here is what you lose, and how to set it up.
Read More
CloudAugust 202613 min read
Claude Code on the Web: Cloud Sessions and Teleport
A cloud session clones your GitHub remote rather than your working copy, so push first. Here is the handoff in both directions, and the part that only goes one way.
Read More
CloudAugust 202613 min read
Claude Code Behind an LLM Gateway: What Breaks and Why
A gateway gives you one place for credentials, budgets and audit logs. It also quietly turns off several Claude Code features, and the failures do not announce themselves.
Read More