"Rolling deletion" is not supported and was never enrolled for customer. So I would recommend using Store OpenTelemetry traces in Unity Catalog (GA'd today) for storing directly in UC which is scalabable beyond the 100K limit. 

The only option is to delete historical traces: https://mlflow.org/docs/latest/genai/tracing/observe-with-traces/delete-traces/ -- you could N traces by using this code snippet

import time
from mlflow import MlflowClient

client = MlflowClient()

# Get current timestamp in milliseconds
current_time = int(time.time() * 1000)

# Delete traces older than current time, limit to 100 traces
deleted_count = client.delete_traces(
    experiment_id="1", max_timestamp_millis=current_time, max_traces=100
)

print(f"Deleted {deleted_count} traces")