Repository navigation
Conversation
…beneath Rate limits counted per key, so a team's throughput scaled with how many keys it minted and the control plane capped key counts to bound it. Now the rows pushed on a key are its team's, counted on lim:team:<team>:<model>:..., and the key's own rows ride beside them as key_limits, counted on lim:key:<prefix>:<model>:...; both gates run in the one admit script and the lower one refuses first. An entry may name a model last (metric:window:value[:model]); a row naming the model stands in for the blank row on its metric and window, and a model no row reaches is uncapped on that list. The model is not known before the body, so Admit leaves quota for modelaccess, after the grant check; a fallback is not admitted again. The 429 names the model and whose limit it was.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Rate limits counted per key, so a team's throughput scaled with how many keys it minted and the control plane capped key counts to bound it. Now the rows pushed on a key are its team's, counted on lim:team:::..., and the key's own rows ride beside them as key_limits, counted on lim:key:::...; both gates run in the one admit script and the lower one refuses first. An entry may name a model last (metric:window:value[:model]); a row naming the model stands in for the blank row on its metric and window, and a model no row reaches is uncapped on that list.
The model is not known before the body, so Admit leaves quota for modelaccess, after the grant check; a fallback is not admitted again. The 429 names the model and whose limit it was.