feat: wire flow control policies and smart queue defaults#2232
Merged
k8s-ci-robot merged 2 commits intokubernetes-sigs:mainfrom Jan 29, 2026
Merged
Conversation
Adds the `FlowControlConfig` API to the `EndpointPickerConfig` CRD and wires it into the Flow Controller initialization path. This change exposes critical operational knobs including: - Global and Per-Band Capacity (`MaxBytes`) - Safety Timeouts (`DefaultRequestTTL`) - Memory Reclamation Tuning (`ExpiryCleanupInterval`)
Adds configuration fields for Fairness and Ordering policies to the `PriorityBandConfig` API and implements intelligent queue selection logic. Key Changes: - API: Added `FairnessPolicyRef` and `OrderingPolicyRef` to Priority Bands, allowing users to select specific policy plugins (e.g., EDF, RoundRobin). - Registry: Implemented "Smart Defaulting" to automatically select `MaxMinHeap` when policies require priority capabilities (e.g., EDF), otherwise defaulting to the more performant `ListQueue` (O(1)). - Loader: Updated the config loader to ensure system-default policy plugins (`fcfs`, `global-strict`) are registered if not explicitly configured by the user. - Plugins: Refactored policy plugins to use explicit Factory registration and support named instances. - Docs: Reorganized the Plugin Configuration guide by category (Infrastructure, Scheduling, Flow Control) for better readability and added content for the new config options.
k8s-ci-robot
added
kind/feature
✅ Deploy Preview for gateway-api-inference-extension ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
k8s-ci-robot
added
size/L
Contributor
Author
LukeAVanDrie
changed the title
Contributor
|
/lgtm |
k8s-ci-robot
added
the
lgtm
Contributor
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: ahg-g, LukeAVanDrie The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
elevran
pushed a commit
to llm-d/llm-d-inference-scheduler
that referenced
this pull request
Apr 23, 2026
…-sigs/gateway-api-inference-extension#2232) * feat: add FC configuration to EndpointPickerConfig Adds the `FlowControlConfig` API to the `EndpointPickerConfig` CRD and wires it into the Flow Controller initialization path. This change exposes critical operational knobs including: - Global and Per-Band Capacity (`MaxBytes`) - Safety Timeouts (`DefaultRequestTTL`) - Memory Reclamation Tuning (`ExpiryCleanupInterval`) * feat: wire FC policies and enable queue defaulting Adds configuration fields for Fairness and Ordering policies to the `PriorityBandConfig` API and implements intelligent queue selection logic. Key Changes: - API: Added `FairnessPolicyRef` and `OrderingPolicyRef` to Priority Bands, allowing users to select specific policy plugins (e.g., EDF, RoundRobin). - Registry: Implemented "Smart Defaulting" to automatically select `MaxMinHeap` when policies require priority capabilities (e.g., EDF), otherwise defaulting to the more performant `ListQueue` (O(1)). - Loader: Updated the config loader to ensure system-default policy plugins (`fcfs`, `global-strict`) are registered if not explicitly configured by the user. - Plugins: Refactored policy plugins to use explicit Factory registration and support named instances. - Docs: Reorganized the Plugin Configuration guide by category (Infrastructure, Scheduling, Flow Control) for better readability and added content for the new config options.
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
4 participants
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.
What type of PR is this?
/kind feature
/kind documentation
/kind cleanup
What this PR does / why we need it:
This PR builds upon #2217 by wiring up the policy plugins. It allows users to configure specific Fairness and Ordering policies (e.g., EDF, Round Robin) per Priority Band and ensures the underlying queue implementation supports the chosen policy's requirements.
Key Changes:
API Extension:
FairnessPolicyRefandOrderingPolicyReftoPriorityBandConfig.Queue Defaulting:
registryto inspect the selected Ordering Policy's capabilities.ListQueue(O(1)) for standard policies like FCFS.MaxMinHeap(O(log n)) if the policy requires heap capabilities (e.g., EDF).Plugin Architecture:
ordering,fairness) to removeinit()side-effects in favor of explicit in-tree factory registration inrunner.go.config/loaderto ensure system-default policy plugins (fcfs-ordering-policy,global-strict-fairness-policy) are automatically registered if not explicitly configured by the user.Documentation Overhaul:
Which issue(s) this PR fixes:
Does this PR introduce a user-facing change?: