Metrics remain after Stream deletion

We are building an application using Redpanda connect.
When running in streams mode, if we delete the file corresponding to a specific stream, that stream is removed, but the associated metrics remain.

As we repeatedly create and delete streams with different names, the unused metrics accumulate, causing the heap to grow. This results in monotonically increasing memory usage.

Is it possible to have the corresponding metrics removed automatically when a stream is deleted?

## log of removing stream 
{"@service":"benthos","level":"info","msg":"Stream trigger-g config deleted, attempting to remove stream."}
{"@service":"benthos","level":"info","msg":"Removed stream trigger-g."}

## trigger-g metrics remain after stream deletion
processor_latency_ns{label="",path="root.pipeline.processors.2.catch.1",stream="trigger-g",quantile="0.5"} NaN
processor_latency_ns{label="",path="root.pipeline.processors.2.catch.1",stream="trigger-g",quantile="0.9"} NaN
processor_latency_ns{label="",path="root.pipeline.processors.2.catch.1",stream="trigger-g",quantile="0.99"} NaN
processor_latency_ns_sum{label="",path="root.pipeline.processors.2.catch.1",stream="trigger-g"} 0
processor_latency_ns_count{label="",path="root.pipeline.processors.2.catch.1",stream="trigger-g"} 0

No source file or test is named in the report. Start by reproducing the issue in streams mode, delete a stream configuration, and trace the removal path alongside the metrics registry. Done means metrics for the deleted stream disappear and repeatedly creating and deleting streams no longer causes unused metrics or monotonic heap growth.