[BUG] force_min_step_interval on victoriametrics backend is ignored

[mlausch]

Describe the bug
The "force_min_step_interval" option is at least in the victoriaMetrics backend ignored.

As I saw on the code, if the value is higher than the interval (stop - start) the step size shold be used for the query on the backend. But this is not the case.

CarbonAPI Version
0.16.0
Also on the Master the bug is present.

CarbonAPI Configuration:

    backendsv2:
        backends:
          - groupName: "victoriametrics"
            protocol: "victoriametrics"
            lbMethod: "broadcast"
            maxTries: 3
            maxBatchSize: 0
            keepAliveInterval: "10s"
            concurrencyLimit: 0
            maxIdleConnsPerHost: 1000
            backendOptions:
                step: "60"
                start: "-43200m"
                max_points_per_query: 5000
                probe_version_interval: "600s"
                fallback_version: "v0.0.0"
                force_min_step_interval: "8800h"
                # vmClusterTenantID: "0" # use vmClisterTenantID for VM-cluster only
            timeouts:
                find: "2s"
                render: "50s"
                connect: "200ms"
            servers:
              - "http://127.0.0.1:8482"
              - "http://127.0.0.1:8483"

Simplified query (if applicable)
query parameter send from grafana to carbonapi:
target=some.metric&from=1730980363&until=1738756365&format=json&maxDataPoints=1378

query parameter send to victoriametrics
end=1738756365&max_lookback=7200s&query=%7B__graphite__%3D%22some.metric%22%7D&start=1730980363&step=7200s

as you can see, the step size is way bigger than the expected "60s" which is the min step size configured in the carbonapi.conf