bug: Adaptative batching with max_batch_size=1 crashes the API

[bruno-hays]

Describe the bug

With this decorator on my function:

    @bentoml.api(batchable=True,
                 max_batch_size=1,
                 max_latency_ms=3600000)

I get this incomprehensible error:

  File "/home/gcpuser/sky_workdir/whisperapi/venv/lib/python3.11/site-packages/bentoml/_internal/utils/metrics.py", line 44, in exponential_buckets
    assert start < end
           ^^^^^^^^^^^
AssertionError

now use this header:

    @bentoml.api(batchable=True,
                 max_batch_size=2,
                 max_latency_ms=3600000)

And it works.

Amusingly, this happens when I run my API on debian and I am not able to reproduce this error on my mac M1

To reproduce

No response

Expected behavior

No response

Environment

bentoml==1.2.19
python==3.11.9