Removed allocation struct used in modelbased scaling#553
Merged
asm582 merged 1 commit intollm-d:mainfrom Jan 9, 2026
Merged
Conversation
| // including the rate of incoming requests (ArrivalRate) and the average | ||
| // length of each request (AvgLength). Both fields are specified as strings | ||
| // to allow flexible input formats. | ||
| type LoadProfile struct { |
Collaborator
There was a problem hiding this comment.
Why do we still need this?
Collaborator
Author
There was a problem hiding this comment.
backward compatibility from an algorithm perspective. I assume we will soon need features for model-based work, either to handle wobble or to go beyond it. If this is absolutely not needed, we can remove it later.
| MaxBatch int `json:"maxBatch"` | ||
|
|
||
| // ITLAverage is the average inter token latency for the current allocation. | ||
| ITLAverage string `json:"itlAverage"` |
Collaborator
There was a problem hiding this comment.
Is this still needed?
Collaborator
Author
There was a problem hiding this comment.
Same comment as above. Until the wobble issue is fixed, I would like to keep it.
| ITLAverage string `json:"itlAverage"` | ||
|
|
||
| // TTFTAverage is the average time to first token for the current allocation | ||
| TTFTAverage string `json:"ttftAverage"` |
Collaborator
There was a problem hiding this comment.
Is this still needed?
Collaborator
Author
There was a problem hiding this comment.
I kept it as a safeguard in case we ever want to bring back model-based features. The user-facing API does not have these fields for now.
lionelvillard
approved these changes
Jan 9, 2026
github-actions bot
pushed a commit
that referenced
this pull request
Jan 9, 2026
Remove documentation for CurrentAlloc, Allocation, and LoadProfile types that were removed in PR #553 as part of the model-based scaling cleanup. Changes: - Remove Allocation and LoadProfile type definitions from CRD reference - Remove currentAlloc field from VariantAutoscalingStatus documentation - Update status examples to use desiredOptimizedAlloc instead - Update VariantAutoscalingStatus description to reflect current state These types were part of the model-based scaling implementation and are no longer needed for saturation-based scaling.
This was referenced Jan 9, 2026
ev-shindin
pushed a commit
to ev-shindin/workload-variant-autoscaler
that referenced
this pull request
Jan 14, 2026
Removed allocation struct used in modelbased scaling
ev-shindin
pushed a commit
to ev-shindin/workload-variant-autoscaler
that referenced
this pull request
Jan 14, 2026
Remove documentation for CurrentAlloc, Allocation, and LoadProfile types that were removed in PR llm-d#553 as part of the model-based scaling cleanup. Changes: - Remove Allocation and LoadProfile type definitions from CRD reference - Remove currentAlloc field from VariantAutoscalingStatus documentation - Update status examples to use desiredOptimizedAlloc instead - Update VariantAutoscalingStatus description to reflect current state These types were part of the model-based scaling implementation and are no longer needed for saturation-based scaling.
mamy-CS
pushed a commit
to mamy-CS/inferno-autoscaler
that referenced
this pull request
Feb 10, 2026
Remove documentation for CurrentAlloc, Allocation, and LoadProfile types that were removed in PR llm-d#553 as part of the model-based scaling cleanup. Changes: - Remove Allocation and LoadProfile type definitions from CRD reference - Remove currentAlloc field from VariantAutoscalingStatus documentation - Update status examples to use desiredOptimizedAlloc instead - Update VariantAutoscalingStatus description to reflect current state These types were part of the model-based scaling implementation and are no longer needed for saturation-based scaling.
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
2 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.
allocation struct has ITL, TTFT, and load average fields. This was handled by model-based scaling, which is not required for saturation-based scaling and therefore is not part of the API.