<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Inside The AI Stack — new resources</title><description>Engineering guides, production infrastructure, hands-on labs, and AI-powered operations tools for the systems behind modern AI.</description><link>https://insidetheaistack.com/</link><language>en-US</language><item><title>AI Agents in Production — Containment, Cost, and Termination</title><link>https://insidetheaistack.com/ai-engineering/ai-agents-in-production/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-engineering/ai-agents-in-production/</guid><description>What changes when an agent loop moves from a demo to production: hard limits, tool authorisation, failure handling, and knowing when an agent is the wrong shape entirely. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>LLM</category><category>Agents</category><category>MCP</category></item><item><title>AI Networking Fundamentals — Why the Fabric Sets the Limit</title><link>https://insidetheaistack.com/ai-infrastructure/ai-networking-fundamentals/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-infrastructure/ai-networking-fundamentals/</guid><description>Collective communication, RDMA, and rail topology explained from the operator&apos;s side, including why adding nodes to a training job can make it slower. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>RDMA</category><category>InfiniBand</category><category>Networking</category><category>NCCL</category></item><item><title>Container Startup Failures — One Guide to the Whole Class</title><link>https://insidetheaistack.com/devops/docker/container-startup-troubleshooting/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/docker/container-startup-troubleshooting/</guid><description>Exec format errors, permission denied, missing commands, bad entrypoints, immediate exits, and CrashLoopBackOff — diagnosed as one family rather than a page per error string. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Docker</category><category>Containers</category><category>Kubernetes</category></item><item><title>Docker Production Engineering — Images That Behave Under an Orchestrator</title><link>https://insidetheaistack.com/devops/docker/docker-production-engineering/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/docker/docker-production-engineering/</guid><description>Building container images for production: reproducible builds, correct signal handling, non-root runtime, layer strategy, and keeping secrets out of image history. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Docker</category><category>Containers</category><category>OCI</category></item><item><title>GPU Infrastructure for AI Workloads — Memory, Bandwidth, and Topology</title><link>https://insidetheaistack.com/ai-infrastructure/gpu-infrastructure-for-ai-workloads/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-infrastructure/gpu-infrastructure-for-ai-workloads/</guid><description>What determines accelerator performance in production: memory capacity versus bandwidth, interconnect topology, and the checks that find a misplaced workload. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>GPU</category><category>NVIDIA</category><category>CUDA</category><category>Kubernetes</category></item><item><title>Incident Diagnosis — Working From Evidence Instead of Intuition</title><link>https://insidetheaistack.com/operations/incident-diagnosis-method/</link><guid isPermaLink="true">https://insidetheaistack.com/operations/incident-diagnosis-method/</guid><description>A repeatable method for diagnosing production failures: form hypotheses the evidence supports, order checks by what they eliminate, and know when to stop and mitigate. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Incident Response</category><category>Observability</category></item><item><title>Kolla-Ansible Production Architecture — Deploying OpenStack in Containers</title><link>https://insidetheaistack.com/cloud/openstack/kolla-ansible-production-architecture/</link><guid isPermaLink="true">https://insidetheaistack.com/cloud/openstack/kolla-ansible-production-architecture/</guid><description>How Kolla-Ansible actually structures an OpenStack deployment, where configuration comes from, and how to make changes without discovering them during an outage. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>OpenStack</category><category>Kolla-Ansible</category><category>Ansible</category><category>Docker</category></item><item><title>Diagnosing Kubernetes Workload Failures — Pending, CrashLoop, and Evicted</title><link>https://insidetheaistack.com/devops/kubernetes/kubernetes-workload-troubleshooting/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/kubernetes/kubernetes-workload-troubleshooting/</guid><description>A single diagnostic path for workloads that will not schedule, will not stay up, or keep getting evicted — driven by events and previous logs rather than guesswork. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Kubernetes</category></item><item><title>OpenStack Production Operations — Reading a Cloud That Is Lying to You</title><link>https://insidetheaistack.com/cloud/openstack/openstack-production-operations/</link><guid isPermaLink="true">https://insidetheaistack.com/cloud/openstack/openstack-production-operations/</guid><description>Operating OpenStack in production: service state versus status, the message bus, placement disagreements, and the quiet failures that keep dashboards green. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>OpenStack</category><category>Nova</category><category>Neutron</category><category>Cinder</category><category>RabbitMQ</category></item><item><title>Production AI Application Architecture — Boundaries That Contain Failure</title><link>https://insidetheaistack.com/ai-engineering/production-ai-application-architecture/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-engineering/production-ai-application-architecture/</guid><description>How to structure an AI application so that a slow model, a failed tool call, or a bad retrieval degrades one part of the system instead of the whole request path. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>LLM</category><category>RAG</category><category>Architecture</category></item><item><title>Production Kubernetes Operations — Resources, Probes, and Rollouts</title><link>https://insidetheaistack.com/devops/kubernetes/production-kubernetes-operations/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/kubernetes/production-kubernetes-operations/</guid><description>The operational settings that decide whether a workload survives a node failure or a rollout: requests and limits, the three probes, disruption budgets, and scheduling. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Kubernetes</category><category>Containers</category></item><item><title>Production Observability — Instrumentation That Answers Questions</title><link>https://insidetheaistack.com/operations/production-observability-engineering/</link><guid isPermaLink="true">https://insidetheaistack.com/operations/production-observability-engineering/</guid><description>Building observability that shortens incidents rather than producing dashboards: what to instrument, how to alert on symptoms, and why cardinality decides your bill. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Prometheus</category><category>Grafana</category><category>OpenTelemetry</category><category>Observability</category></item><item><title>RAG Architecture — Chunking, Retrieval, and Why the Model Is Rarely at Fault</title><link>https://insidetheaistack.com/ai-engineering/rag-architecture-guide/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-engineering/rag-architecture-guide/</guid><description>A working guide to retrieval-augmented generation: how to chunk, how to evaluate retrieval separately from generation, and how to diagnose a bad answer in the right order. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>RAG</category><category>Embeddings</category><category>LLM</category></item><item><title>Storage Architecture for AI — Three Access Patterns, Three Failure Modes</title><link>https://insidetheaistack.com/ai-infrastructure/storage-architecture-for-ai/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-infrastructure/storage-architecture-for-ai/</guid><description>Dataset streaming, checkpoint bursts, and model loading place completely different demands on storage. Designing for one and getting the others wrong is the usual outcome. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Storage</category><category>NVMe</category><category>Ceph</category><category>AI Infrastructure</category></item><item><title>Terraform Production Practices — Plan Review, State, and Blast Radius</title><link>https://insidetheaistack.com/devops/terraform/terraform-production-practices/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/terraform/terraform-production-practices/</guid><description>Running Terraform against production infrastructure: reviewing the plan mechanically, containing blast radius, structuring state, and avoiding the destroys nobody noticed. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Terraform</category><category>OpenTofu</category><category>IaC</category></item><item><title>Terraform State Operations — Import, Move, and Recover Without Losing Anything</title><link>https://insidetheaistack.com/devops/terraform/terraform-state-operations/</link><guid isPermaLink="true">https://insidetheaistack.com/devops/terraform/terraform-state-operations/</guid><description>The state commands that are genuinely dangerous, what each one actually does, and how to run them with a way back. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>Terraform</category><category>OpenTofu</category><category>IaC</category></item><item><title>The AI Stack Explained — From Application Down to Silicon</title><link>https://insidetheaistack.com/ai-engineering/the-ai-stack-explained/</link><guid isPermaLink="true">https://insidetheaistack.com/ai-engineering/the-ai-stack-explained/</guid><description>A layer-by-layer map of the modern AI stack, what each layer is actually responsible for, and where production systems tend to break in practice. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>articles</category><category>LLM</category><category>Kubernetes</category><category>GPU</category><category>RAG</category></item><item><title>Cinder Volume Creation Fails With &quot;No Valid Backend Was Found&quot;</title><link>https://insidetheaistack.com/runbooks/openstack/cinder-scheduler-no-valid-backend/</link><guid isPermaLink="true">https://insidetheaistack.com/runbooks/openstack/cinder-scheduler-no-valid-backend/</guid><description>Diagnose and remediate OpenStack volume creation failures caused by the scheduler having no candidate backends, including the case where the API reports healthy. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>runbooks</category><category>OpenStack</category><category>Cinder</category><category>Ceph</category></item><item><title>Triaging a Pod in CrashLoopBackOff</title><link>https://insidetheaistack.com/runbooks/kubernetes/kubernetes-crashloopbackoff-triage/</link><guid isPermaLink="true">https://insidetheaistack.com/runbooks/kubernetes/kubernetes-crashloopbackoff-triage/</guid><description>A structured procedure for a pod that will not stay running, covering exit-code classification, probe-induced restarts, OOM kills, and validation that the fix held. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>runbooks</category><category>Kubernetes</category></item><item><title>Responding to Disk Pressure on a Linux Host</title><link>https://insidetheaistack.com/runbooks/linux/linux-disk-pressure-response/</link><guid isPermaLink="true">https://insidetheaistack.com/runbooks/linux/linux-disk-pressure-response/</guid><description>Find what is consuming a full filesystem, reclaim space safely including the deleted-but-open-file case, and validate before the host causes wider failures. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>runbooks</category><category>Linux</category><category>Kubernetes</category></item><item><title>Lab: A Deployment That Will Not Stay Up</title><link>https://insidetheaistack.com/labs/kubernetes-crashloopbackoff/</link><guid isPermaLink="true">https://insidetheaistack.com/labs/kubernetes-crashloopbackoff/</guid><description>A service crash-looping after a routine config change, with logs that show a normal startup and no error anywhere. Work out what is killing it. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>labs</category><category>Kubernetes</category></item><item><title>Lab: Diagnose a Nova Compute Failure Returning NoValidHost</title><link>https://insidetheaistack.com/labs/openstack-nova-novalidhost/</link><guid isPermaLink="true">https://insidetheaistack.com/labs/openstack-nova-novalidhost/</guid><description>Instances stopped building on a cloud with visible free capacity. Work the evidence, form hypotheses, and find why the scheduler has no candidates. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>labs</category><category>OpenStack</category><category>Nova</category><category>Placement</category></item><item><title>Lab: The Plan That Would Have Deleted Production</title><link>https://insidetheaistack.com/labs/terraform-destructive-plan-review/</link><guid isPermaLink="true">https://insidetheaistack.com/labs/terraform-destructive-plan-review/</guid><description>A routine-looking Terraform pull request. Find the changes that would destroy data before you approve it, and work out which one is not what it appears to be. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>labs</category><category>Terraform</category><category>AWS</category></item><item><title>OpenStack Production Engineer</title><link>https://insidetheaistack.com/academy/openstack-production-engineer/</link><guid isPermaLink="true">https://insidetheaistack.com/academy/openstack-production-engineer/</guid><description>A sequenced learning path for engineers who operate OpenStack in production — architecture, service-by-service depth, deployment, and troubleshooting under pressure. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>tracks</category><category>OpenStack</category><category>Nova</category><category>Neutron</category><category>Cinder</category><category>Kolla-Ansible</category></item><item><title>The OpenStack Control Plane — Who Talks to Whom</title><link>https://insidetheaistack.com/academy/openstack-production-engineer/openstack-control-plane-architecture/</link><guid isPermaLink="true">https://insidetheaistack.com/academy/openstack-production-engineer/openstack-control-plane-architecture/</guid><description>How OpenStack services communicate, why the message bus and database are the real dependencies, and what each service actually owns. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>lessons</category><category>OpenStack</category><category>RabbitMQ</category><category>MariaDB</category></item><item><title>Kolla-Ansible Configuration — Making a Change That Survives</title><link>https://insidetheaistack.com/academy/openstack-production-engineer/openstack-kolla-configuration/</link><guid isPermaLink="true">https://insidetheaistack.com/academy/openstack-production-engineer/openstack-kolla-configuration/</guid><description>Where OpenStack configuration actually comes from in a Kolla-Ansible deployment, how overrides are merged, and how to scope a change so it does not touch the whole cloud. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>lessons</category><category>OpenStack</category><category>Kolla-Ansible</category><category>Ansible</category></item><item><title>Nova Scheduling — Placement, Filters, and Reading a Rejection</title><link>https://insidetheaistack.com/academy/openstack-production-engineer/openstack-nova-scheduling/</link><guid isPermaLink="true">https://insidetheaistack.com/academy/openstack-production-engineer/openstack-nova-scheduling/</guid><description>How Nova decides where an instance runs, what the filter chain does, and how to find which filter eliminated every host. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>lessons</category><category>OpenStack</category><category>Nova</category><category>Placement</category></item><item><title>Service State and Liveness — Reading What the Cloud Is Telling You</title><link>https://insidetheaistack.com/academy/openstack-production-engineer/openstack-service-state-and-liveness/</link><guid isPermaLink="true">https://insidetheaistack.com/academy/openstack-production-engineer/openstack-service-state-and-liveness/</guid><description>Status versus state, how liveness is actually determined, and why a service that is enabled and down is the most dangerous row in any OpenStack listing. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>lessons</category><category>OpenStack</category><category>Nova</category><category>Neutron</category></item><item><title>AI Infrastructure Operations Survey — 2027 Edition</title><link>https://insidetheaistack.com/research/ai-infrastructure-operations-survey/</link><guid isPermaLink="true">https://insidetheaistack.com/research/ai-infrastructure-operations-survey/</guid><description>A planned survey of how teams actually operate AI infrastructure in production — what breaks, what is measured, and where engineering time goes. Not yet conducted. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>research</category><category>AI Infrastructure</category><category>GPU</category></item><item><title>GPU Cloud Price Index — Methodology</title><link>https://insidetheaistack.com/research/gpu-cloud-price-index/</link><guid isPermaLink="true">https://insidetheaistack.com/research/gpu-cloud-price-index/</guid><description>A planned recurring measurement of accelerator pricing across cloud providers, normalised so the comparison is meaningful. Methodology published before data collection begins. [Awaiting technical review]</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>research</category><category>GPU</category><category>Cloud</category><category>AI Infrastructure</category></item></channel></rss>