Fix heartbeat races on event updates (elastic#6950) (elastic#7009)

Heartbeat has a many components, each adding fields to an existing
event. As events are created per 'TaskRun', this is normally fine. But
in case the TaskRunner branches of into multiple Sub-tasks, we will see
races on shared event structures. This is the case with multiple ports
in the http/tcp configuration or the IPAll setting (ping all known IPs
of a given domain).

Namespaces in an event can be potentially populated via globally shared
structures. This can also lead to unwanted updates or races.

This fix clones the event structure before continuing event updates, so
to remove the chance of races.

(cherry picked from commit bc225fb)

CHANGELOG.asciidoc

@@ -67,6 +67,7 @@ https://github.com/elastic/beats/compare/v6.0.0-beta2...master[Check the HEAD di
 - Add raw JSON to message field when JSON parsing fails. {issue}6516[6516]
 
 *Heartbeat*
+- Fix race due to updates of shared a map, that was not supposed to be shared between multiple go-routines. {issue}6616[6616]
 
 *Metricbeat*
 

heartbeat/monitors/util.go

@@ -122,6 +122,8 @@ func annotated(
 		}
 
 		if fields != nil {
+			fields = fields.Clone()
+
 			status := look.Status(err)
 			fields.DeepUpdate(common.MapStr{
 				"monitor": common.MapStr{
@@ -130,7 +132,7 @@ func annotated(
 				},
 			})
 			if user := settings.Fields; user != nil {
-				fields.DeepUpdate(user)
+				fields.DeepUpdate(user.Clone())
 			}
 
 			event.Timestamp = start
@@ -370,10 +372,13 @@ func resolveErr(host string, err error) (common.MapStr, []TaskRunner, error) {
 func WithFields(fields common.MapStr, r TaskRunner) TaskRunner {
 	return MakeCont(func() (common.MapStr, []TaskRunner, error) {
 		event, cont, err := r.Run()
-		if event == nil {
+		if event != nil {
+			event = event.Clone()
+			event.DeepUpdate(fields)
+		} else if err != nil {
 			event = common.MapStr{}
+			event.DeepUpdate(fields)
 		}
-		event.DeepUpdate(fields)
 
 		for i := range cont {
 			cont[i] = WithFields(fields, cont[i])