Cost spike playbooks for every connected provider.
Common causes, first checks, monitoring pages, setup guides, and alert policy templates for cloud, AI, data, developer tooling, and communications spend.
Diagnose the spike
Start with the failure modes that most often move spend for each provider.
Set the policy
Use green, amber, and red thresholds your team can review every morning.
Monitor daily
Link straight into StackSpend monitoring and setup pages from each playbook.
Cloud cost spike guides
AWS
AWS spikes are usually caused by resources that keep running, data transfer that compounds quietly, or managed services that scale faster than the team reviews billing.
GCP
GCP cost jumps often come from BigQuery query volume, autoscaling compute, project sprawl, or billing exports that no one reviews daily.
Azure
Azure spend tends to spike when reservations lapse, workloads autoscale, or subscription-level reporting hides the service that changed.
Vercel
Vercel costs often spike when traffic, image transformations, serverless functions, or build activity scale faster than the release plan.
AI cost spike guides
OpenAI
OpenAI spend can move in hours because token volume, model choice, retries, and product traffic all multiply together.
Anthropic
Anthropic spend often jumps when long-context workflows, agents, or premium Claude models become part of a high-volume path.
Cursor
Cursor costs usually rise through seat growth, heavier engineering usage, or team-wide coding-agent workflows that become normal overnight.
Hugging Face
Hugging Face spend usually rises when GPU-backed endpoints, Spaces, or Jobs are left running after experiments become infrastructure.
Claude
Claude usage spikes when internal tools, product features, or agentic workflows create more model calls than the team expects.
Grok
Grok spend can increase quickly when experiments, model routing, or customer-facing AI features move into higher-volume paths.