LS constantly reloading failing pipeline & exhausting host connections

Noticed that when I have a pipeline, with an Elasticsearch output, that is failing, LS is then reloading the pipeline over & over but not clearing down the previously spawned connections to ES so I end up with connection limits being hit .

Using pipeline.yml to load in individual pipeline configs.

Is this a known issue or is there a config entry i can use to stop the connection exhaustion