AI GlossaryㅊInfrastructure and chips
dimension keys
The criteria used to divide AI traffic into different groups when applying rate limits
In plain words
Dimension keys are the criteria used to split AI traffic into separate groups. Think of an amusement park that sorts visitors into a general line and a fast-track line based on which ticket they hold. Here, 'which ticket someone holds' is the dimension key. Just as the ticket type decides which line you stand in, things like user group or the model being requested decide which processing path traffic gets routed through.
Each group created this way gets its own rules for how fast and how much it can process. Just as a beginner-ticket line moves slowly while a premium line moves quickly, groups split by dimension keys can each have different rate limits applied. This lets administrators fine-tune limits instead of applying the same cap to everyone — for example, giving a looser limit only to a group piloting a new feature.
How it shows up in the news
Articles describe it this way: 'A rate-limiting rule consists of dimension keys, which group requests, and entries, which define the throughput allowed for each group.' Here, dimension keys are simply the criteria for grouping by things like user or model — they are not the rate limit values themselves. The actual throughput caps are set in the entries.
See also
Stories using this term
- AWS adds AI traffic rate limiting to AgentCore gatewayAI · 2026.08.09
- AWS unveils build guide for automated web insight extraction with Bedrock AgentCoreAI · 2026.08.09
- AWS unveils bridge letting cloud agents use local MCP toolsAI · 2026.08.09
- AWS unveils temporal policies to verify agent action historyAI · 2026.08.09
- Mobileye Cuts Ticket Handling Time 90% With AI AgentBusiness · 2026.08.06
- AWS Adds Open-Source Agent Skills for Bedrock Automated Reasoning PoliciesAI · 2026.08.09
