SseActor failing on connection timeout

Gatling (Open-Source)
1

I’m not sure this is expected behaviour, but when we overload our test SSE server, Gatling starts spewing errors.
This seems to cause the test to hang waiting for the failed clients to complete.

Gatling 2.2.4 running on OSX. Java 1.8.0_144 (OSX)

11:21:07.920 [ERROR] a.a.OneForOneStrategy - Actor io.gatling.http.action.async.sse.SseActor@1843d3d5 doesn’t support message OnThrowable(AsyncTx(Session(RefMan client SSE connection,8143,Map(gatling.http.cache.dns → io.gatling.http.resolver.ShuffleJdkNameResolver@28086a54, unique_id → 44c665bb-3025-3929-b216-68cdfd9104da, authHeader → Bearer foo, id → 11359037401, profile_id → 29447751, profile_uuid → 4875739e-2c71-37f8-beed-365af1025538),1511781657817,0,OK,List(),io.gatling.core.protocol.ProtocolComponentsRegistry$$Lambda$186/1742900195@37385567),io.gatling.core.action.Pause@78293a49,notifications,https://notifications.example.com/sse GET headers: Accept:text/event-stream
Cache-Control:no-cache Authorization:Bearer foo,HttpProtocol(List(https://notifications-staging.mendeley.com),Some(http://gatling.io),HttpProtocolEnginePart(true,false,6,false,Map(),None,List()),HttpProtocolRequestPart(Map(),None,true,true,false,None,false,None),HttpProtocolResponsePart(true,None,false,true,None,List(),None,false,None),HttpProtocolWsPart(List(),false,None),HttpProtocolProxyPart(None,List())),1511781657817,0,Some(AsyncCheck(CheckBase(io.gatling.http.check.async.AsyncCheckBuilders$$$Lambda$90/1639765351@e3f9e06,,,None),true,60 seconds,UntilCount(1),1511781657817)),List(),List()),io.gatling.http.action.async.sse.SseHandler$$anon$1: Server returned http response with code 504,1511781667886)

The relevant part of the scenario setup is

feed(xxx).exec(
  sse("notifications")
    .open(**"https://notifications.example.com/foo"**)
    .header("Authorization", "${authHeader}")
)
.pause(200 seconds)
.exec(sse("Close SSE").close())






thanks

  Ash