Multisearch ignores memory circuit breaker errors

[mvdz]

Demonstrated by this code which temporarily sets the cluster limit to 1 byte in order to force the error:

import com.sksamuel.elastic4s.ElasticDsl._
import com.sksamuel.elastic4s.http.JavaClient
import com.sksamuel.elastic4s.{ElasticClient, ElasticProperties}

val client = ElasticClient(JavaClient(ElasticProperties("http://localhost:9200")))
println(client.execute(clusterTransientSettings(Map("indices.breaker.total.limit" -> "1b"))).await)
val request = multi(search("_all"))
val response = client.execute(request).await
println(response)
println(client.execute(clusterTransientSettings(Map("indices.breaker.total.limit" -> null))).await)
client.close()

Actual result: RequestSuccess response with an empty list of items:

RequestSuccess(
  429,
  Some({"error":{"root_cause":[{"type":"circuit_breaking_exception","reason":"[parent] Data too large, data for [<http_request>] would be [220036656/209.8mb], which is larger than the limit of [1/1b], real usage: [220036616/209.8mb], new bytes reserved: [40/40b], usages [model_inference=0/0b, eql_sequence=0/0b, fielddata=0/0b, request=0/0b, inflight_requests=40/40b]","bytes_wanted":220036656,"bytes_limit":1,"durability":"TRANSIENT"}],"type":"circuit_breaking_exception","reason":"[parent] Data too large, data for [<http_request>] would be [220036656/209.8mb], which is larger than the limit of [1/1b], real usage: [220036616/209.8mb], new bytes reserved: [40/40b], usages [model_inference=0/0b, eql_sequence=0/0b, fielddata=0/0b, request=0/0b, inflight_requests=40/40b]","bytes_wanted":220036656,"bytes_limit":1,"durability":"TRANSIENT"},"status":429}),
  Map(X-elastic-product -> Elasticsearch, content-type -> application/json, content-length -> 849),
  MultiSearchResponse(List())
)

Expected result is a: RequestFailure(429,...

Tested with the latest releases: elastic4s 8.9.2 and Elasticsearch 8.10.2