regression: query_frontend.metrics.max_duration is inaccurate after upgrading to 2.10.1

[ghudiczius-ahold]

Describe the bug
After upgrading tempo from 2.10.0 to 2.10.1 (via upgrading the helm chart from 2.1.2 to 2.1.3) our test suite started failing with the following error when trying to execute metrics query at tempo:

curl: (22) The requested URL returned error: 400
metrics query time range exceeds the maximum allowed duration of 8h0m0s

Our related configuration is query_frontend.metrics.max_duration: 8h overriding the default 3h (query_frontend.config.metrics.max_duration: 8h in the helm values)

To Reproduce
Steps to reproduce the behavior:

1. Start Tempo (2.10.1) with query_frontend.metrics.max_duration set to 8h
2. curl --silent --show-error --fail-with-body --request GET tempo/api/metrics/query_range --data-urlencode 'q={} | rate()' --data-urlencode 'start=1772080185' --data-urlencode 'end=1772108985' --get

Expected behavior
HTTP/200 with a JSON response

Environment:

• Infrastructure: kubernetes
• Deployment tool: helm

Additional Context

Tested with the default value for query_frontend.metrics.max_duration (3h), and seen the same issue

The issues is also reproducible via grafana