YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
HPA Rightsizing Recommendations
Model card capturing the CPU rightsizing recommendation pass for the apps namespace.
Policy
Sourced from the rightsizing policy dataset (rightsizing-policy, file policy.jsonl):
target_util_pct: 70low_watermark_pct: 40high_watermark_pct: 85
Method
For each Deployment in namespace apps:
util_pct = round(100 * used / request)- If
util_pct > 85->scale_up;rec_millicores = ceil(used / (target/100))rounded up to next multiple of 50. - Else if
util_pct < 40->scale_down; same formula. - Otherwise ->
keep;rec_millicores= current request.
Summary
Total recommended CPU: 3150 millicores Deployments to scale up: 2
Recommendations
| deployment | request | used | util_pct | action | rec_millicores |
|---|---|---|---|---|---|
| api-gateway | 1000 | 900 | 90 | scale_up | 1300 |
| billing | 500 | 120 | 24 | scale_down | 200 |
| notifier | 400 | 260 | 65 | keep | 400 |
| reporter | 600 | 90 | 15 | scale_down | 150 |
| scheduler | 800 | 760 | 95 | scale_up | 1100 |
See recommendations.jsonl for the machine-readable per-deployment records.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support