Autoscaling
Autoscaling is the automatic adjustment of computing resources or service capacity in response to workload demand or system conditions.
Kubernetes for Developers
Cluster autoscaling is the automated adjustment of cluster node capacity to provide resources for scheduled workloads.
- How it is used
- Adds provider-managed node capacity when Pods need space, including during workload scaling or a new deployment.
- How deep it goes
- Explains provider setup choices, node-type limits, startup delay and a placeholder-Pod remedy; defers uncovered scheduling cases to provider documentation.
- Examples
- Uses GKE Autopilot and node pools for setup choices, then a low-priority placeholder Deployment to demonstrate spare capacity.
- What you can do after
- Operate/configure it through provider-supported autoscaling, and reserve capacity with placeholder Pods when startup delay matters.
Microsoft Azure in Action
The act of decreasing a virtual machine's size to provide fewer computing resources.
- How it is used
- Matches database throughput or compute capacity to changing traffic while limiting idle capacity during quieter periods.
- How deep it goes
- Explains scaling directions, thresholds, instance limits, and Cosmos DB throughput caps; does not detail the underlying scaling mechanism.
- Examples
- Uses holiday shopping traffic, a Cosmos DB Azure CLI command, and database demand during major sales.
- What you can do after
- Operate/configure it by setting Cosmos DB autoscale and an explicit maximum throughput, while accounting for possible cost spikes.
Acing the Frontend Interview
Autoscaling automatically adjusts the number or capacity of computing resources in response to workload demand or operational metrics.
- How it is used
- Frames autoscaling as an interview-ready explanation of elastic cloud infrastructure and workload response.
- How deep it goes
- Defines scale-out and scale-in, names common metrics, and reaches a Kubernetes HPA manifest without detailing policy tuning.
- Examples
- Uses product launches, holiday sales, high-traffic APIs, and a multi-metric Kubernetes HorizontalPodAutoscaler manifest.
- What you can do after
- Operate/configure autoscaling by selecting metrics and discussing replica limits for a Kubernetes-hosted API.
Read how Acing the Frontend Interview explains autoscaling →
MongoDB 8.0 in Action, Third Edition
Scaling down is reducing a deployment's provisioned cluster tier or resource capacity.
- How it is used
- Manages Atlas cluster tiers and storage capacity in response to workload use while controlling operating costs.
- How deep it goes
- Explains configuration ranges, CPU and memory triggers, cluster-class limits, defaults, and rolling operation, but not custom implementation scripts.
- Examples
- Uses an M30-to-M50 configuration, a memory-use formula, and differing UI, API, and CLI defaults.
- What you can do after
- Operate/configure Atlas autoscaling by setting bounded cluster tiers, storage behavior, and permission to scale down.
Read how MongoDB 8.0 in Action, Third Edition explains autoscaling →
Acing the Certified Kubernetes Administrator Exam, Second Edition
Autoscaling automatically adjusts workload or resource capacity in response to demand or measured conditions.
- How it is used
- Deploys autoscaling as a Kubernetes workload-management task supporting application availability.
- How deep it goes
- Identifies autoscaling as a configuration competency but does not provide implementation commands, manifests, algorithms, or verification procedures.
- Examples
- Uses a competency table that maps configuring workload autoscaling to the chapter’s workload operations.
- What you can do after
- Operate/configure workload autoscaling as a Kubernetes administration and CKA exam task.
Read how Acing the Certified Kubernetes Administrator Exam, Second Edition explains autoscaling →
Google Anthos in Action
Scaling is the adjustment of computing resources or system capacity to handle changing workload demand.
- How it is used
- Deploys serverless workloads and manages cluster capacity in response to demand, resource requirements, or operational constraints.
- How deep it goes
- Explains request-based scaling, scale-to-zero, tuning parameters, and cluster trade-offs without detailing autoscaler internals.
- Examples
- Uses Cloud Run request scaling, Knative instance limits, Anthos on Azure cluster autoscaling, and horizontal or vertical cluster layouts.
- What you can do after
- Operate or configure request-driven services by selecting concurrency and instance bounds while evaluating broader cluster scaling arrangements.
LLMs in Production
Scaling down is the deliberate reduction of active service capacity or computing resources.
- How it is used
- Manages cluster capacity for deployment demand while limiting available resources and avoiding unnecessary fixed infrastructure.
- How deep it goes
- Explains capacity limits and redeployment risks, but stops short of detailing scale-down algorithms or controller behavior.
- Examples
- Uses cloud cluster commands and a configuration listing that caps CPU, memory, T4, A100, and A100 80GB resources.
- What you can do after
- Operate/configure node auto-provisioning with explicit resource caps and treat LLM capacity reductions as potentially slow to reverse.
Bootstrapping Microservices, Second Edition: With Docker, Kubernetes, GitHub Actions, and Terraform
Elastic scaling is the automatic expansion and contraction of a service's running capacity in response to changing workload.
- How it is used
- Automates cluster and microservice capacity changes to match variable demand while reducing unnecessary resource costs.
- How deep it goes
- Explains cluster and replica bounds with Terraform and Kubernetes resources, while deferring detailed autoscaler customization to documentation.
- Examples
- Uses Terraform node-pool settings and an autoscaling/v2 gateway HorizontalPodAutoscaler with minimum and maximum replica counts.
- What you can do after
- Operate/configure Kubernetes autoscaling for cluster nodes and individual microservice replicas.
Designing AI Systems
Auto-scaling is the automatic adjustment of computing resources in response to workload demand.
- How it is used
- Operates workflow deployments by adjusting container replicas in response to metrics and request demand.
- How deep it goes
- Explains replica floors, ceilings, CPU targets, and statelessness, but defers autoscaler implementation internals.
- Examples
- Uses Claw's two-to-ten replica range, 70% CPU target, and independently scaled parallel research workflows.
- What you can do after
- Operate/configure workflow deployments with replica bounds, resource targets, and externalized state for horizontal scaling.
Also covered in
- GitHub Actions in Action: Continuous integration and delivery for DevOps
- Designing Deep Learning Systems: A guide for software engineers
- Distributed Machine Learning Patterns
- Server-Side WebAssembly
- Cloud Observability in Action
- 100 SQL Server Mistakes and How to Avoid Them
- The Cloud at Your Service: The when, how, and why of enterprise cloud computing
- Shipping Go: Develop, deliver, discuss, design, and go again
- Akka in Action, Second Edition
- Think Distributed Systems
- Architecting an Apache Iceberg Lakehouse
- Data-Driven Cybersecurity
What to know first
- containerA software container is an isolated execution environment that packages an application and its dependencies, typically sharing the host operating-system kernel.
- scalabilityScalability is the ability of a system or process to accommodate growth in workload, size, or demand while maintaining acceptable operation.
- costCost is the monetary expenditure or other resource burden incurred to acquire, produce, operate, or perform something.
- CPUA central processing unit (CPU) is a computer processor that executes program instructions and performs general-purpose computational operations.
Related concepts
- HorizontalPodAutoscalerimplemented by · A HorizontalPodAutoscaler is a Kubernetes API resource and controller that adjusts a workload’s replica count according to observed metrics and configured targets.
- horizontal scalingincludes · Horizontal scaling is increasing or decreasing system capacity by adding or removing computing nodes or service instances that share the workload.
- vertical scalingincludes · Vertical scaling increases or decreases a system’s capacity by changing the resources available to an existing machine, node, or instance.
- costproduces · Cost is the monetary expenditure or other resource burden incurred to acquire, produce, operate, or perform something.
- computer clusterconfigures · A computer cluster is a group of interconnected computing nodes coordinated as a shared system to provide resources, execute workloads, or deliver services.
- Kubernetesused by · Kubernetes is an open-source platform for orchestrating containerized workloads and automating their deployment, scaling, and management.
Questions about autoscaling
What is autoscaling?
Autoscaling is the automatic adjustment of computing resources or service capacity in response to workload demand or system conditions.
Which Manning books explain autoscaling?
Kubernetes for Developers, Microsoft Azure in Action, Acing the Frontend Interview, MongoDB 8.0 in Action, Third Edition, Acing the Certified Kubernetes Administrator Exam, Second Edition, Google Anthos in Action, LLMs in Production, Bootstrapping Microservices, Second Edition: With Docker, Kubernetes, GitHub Actions, and Terraform and 1 more.
What should I know before learning autoscaling?
Manning books that teach autoscaling usually introduce container, scalability, cost and CPU first.
What kinds of autoscaling are there?
Kinds of autoscaling include horizontal scaling and vertical scaling.
What is autoscaling also called?
Autoscaling is also called auto-scaling, elastic scaling, automatic scaling or autoscale.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.








