Throttling is the deliberate limiting or controlling of a rate - whether of data, requests, processes, or resources - to prevent overload, ensure fair usage, or maintain system stability.
The term applies across several domains:
Networking / APIs
Rate-limiting the number of requests a client can make in a given time window (e.g., 1000 API calls per minute). Excess requests are delayed, queued, or rejected with a 429 Too Many Requests response.
Computing / CPUs
Reducing a processor's clock speed when it overheats or when power needs to be conserved. The CPU intentionally runs slower to stay within thermal or power limits.
Databases
Limiting query throughput or connection rates to prevent a single user or process from monopolizing resources.
Cloud Services
Providers (AWS, Azure, GCP) throttle service calls to protect infrastructure and enforce service-tier limits.
Engine / Mechanical
In vehicles and engines, the throttle controls airflow and fuel delivery - "throttling" means restricting that flow to reduce power output.
General principle across all uses:
Throttling = intentionally capping a rate or flow to stay within defined limits.
The goal is always the same: balance load, protect stability, and ensure fair access - rather than letting any single consumer exhaust a shared resource.