This issue is related to elastic/kibana#116850 . In that case the node subprocess exited early with a status of 1. This was determined by searching for relevant cmd/status events in Elasticsearch. It turns out we only create a summary doc on a journey/end event as seen in enrich.go. This is not correct, we should always create a summary doc when the monitor is done.

This is a bug in that it's hard for a user to know a check failed without a summary doc. The summary doc should, of course, have an error field matching the cmd/status error field.