AWS bill shock / RunPlacement framework

Cloud Bill Shock Taxonomy

Direct answer: Classify bill shock by driver class first: compute, network, storage, observability, managed services, support, marketplace, or commitment mismatch.

Decision rule
  • Do not migrate because of a surprise bill until the recurring driver class is known.
  • Use provider pricing pages and your own bill or quote before making a purchase or migration decision.
By Andrew Cooper, Founder of RunPlacement Updated May 2026 Provider-neutral, estimate-labeled guidance Verify current provider pricing

Definition

cloud bill shock

Cloud bill shock is a material cloud spend increase that the team cannot immediately explain by expected workload growth or a planned architecture change.

Bill shock delta = current period spend - last normal baseline; recurring shock = delta expected to repeat.
Infographic showing the Cloud Bill Shock Taxonomy: compute, network, storage, observability, managed services, support, and marketplace drivers, followed by a four-step diagnostic flow.
Cloud bill shock gets easier to diagnose when the largest month-over-month delta is classified before changing infrastructure.

Simple version

Short version

Classify bill shock by driver class first: compute, network, storage, observability, managed services, support, marketplace, or commitment mismatch.

Do not migrate because of a surprise bill until the recurring driver class is known.

RunPlacement quiz

Pressure-test this workload

Do not migrate because of a surprise bill until the recurring driver class is known.

Uses workload type, budget, GPU need, data movement, priority, and ops tolerance.
Use the quiz

Example scenarios

Network shock

NAT Gateway, cross-AZ traffic, internet egress, or inter-region movement changed.

Observability shock

Log volume, retention, custom metrics, or debug settings expanded.

Storage shock

Snapshots, retrieval, replication, requests, or lifecycle gaps became visible.

Decision Table

OptionBest useRisk
ComputeIdle instances, autoscaling, GPU, commitmentsResize, delete, reserve, or re-place
NetworkNAT, egress, cross-zone, inter-regionChange route, endpoint, region, or architecture
StorageSnapshots, retrieval, replication, requestsLifecycle, tiering, retention, or access changes
ObservabilityLogs, metrics, traces, retentionRetention, sampling, log level, or vendor review
Managed servicesDatabase, queues, search, analyticsCapacity, replicas, backups, or architecture review

Quality guide

How to use this framework

RunPlacement pages use public provider documentation, source-linked pricing pages where relevant, estimate-labeled examples, and practical decision frameworks. Estimates are directional and should be verified against provider pricing pages before buying or migrating.

Who this is for

  • Teams staring at an AWS or cloud bill spike before they know what changed.
  • Founders deciding whether a surprise bill means they should leave a major cloud.
  • FinOps or engineering leads who need a shared vocabulary for surprise spend.

How to use it

  • Find the largest month-over-month delta by service, region, and account.
  • Classify the driver before proposing migration, vendor changes, or tooling.
  • Use the taxonomy to decide whether the fix is rightsizing, routing, retention, architecture, or placement.

Common mistakes

  • Assuming compute caused the spike before checking network, logs, storage, or managed services.
  • Starting a migration conversation before identifying the bill driver.
  • Averaging the whole bill instead of isolating the largest delta.

When it does not apply

  • Use cloud billing exports or provider cost tools for exact line-item accounting.
  • Use incident review when the bill spike came from an outage or abuse event.
  • Use procurement review when support or marketplace contracts changed.

Worked examples and scenarios

NAT Gateway spike

A workload may not be compute-heavy at all. Private subnet traffic, cross-AZ paths, and managed-service routes can create a networking bill that looks like general cloud inflation.

Log volume spike

Debug logging, trace retention, or a noisy service can turn observability into the largest delta. Fix retention and sampling before changing providers.

Managed database jump

Replicas, provisioned throughput, backups, or storage growth can create bill shock without a placement problem. The taxonomy keeps the fix close to the driver.

Related decisions

Apply the framework

Use these long-tail decision pages when a specific cost driver or provider choice is already visible.

Related resources

Turn the framework into a worksheet

These checklists make the concept easier to share and apply.

FAQ

What should I check first after cloud bill shock?

After cloud bill shock, check the largest month-over-month delta by service, region, and account before changing architecture. Then classify the driver as compute, network, storage, observability, managed service, support, marketplace, or commitment mismatch. The recurring driver should determine the next action.

Is cloud bill shock always a migration signal?

No. Cloud bill shock is often a cleanup, retention, routing, idle resource, or architecture issue inside the current cloud. Migration only becomes a serious option after the recurring driver is understood and the expected savings beat migration work, downtime risk, and new operations.

What is a recurring bill shock?

A recurring bill shock is the part of the cost increase expected to repeat if nothing changes. Separate it from experiments, one-time data movement, temporary traffic, or project spikes. Recurring deltas matter most because they can justify optimization, architecture changes, commitments, or migration analysis.

Sources

RunPlacement quiz

Pressure-test this workload

Do not migrate because of a surprise bill until the recurring driver class is known.

Uses workload type, budget, GPU need, data movement, priority, and ops tolerance.
Use the quiz