Test our microbenchmarks as part of CI

### Is your feature request related to a problem or challenge?

We can test our benchmarks with `cargo test --benches` which runs in debug & single iteration, just to ensure the benchmark code actually executes. Have seen instances in arrow-rs where some were panicking but we didn't know since we didn't often run those benchmarks

### Describe the solution you'd like

_No response_

### Describe alternatives you've considered

_No response_

### Additional context

- Potentially related to https://github.com/apache/datafusion/issues/15511

Run `cargo test --benches` locally to confirm the intended check, then inspect the repository's existing CI workflows to find the appropriate job. Done means CI executes the microbenchmarks in debug, single-iteration test mode and fails when benchmark code panics; issue #15511 provides related context.

Written by the indexing model from the issue text.