Kibana version : 7.5.1
Elasticsearch version : 7.5.1
APM Server version : 7.5.1
APM Agent language and version : Node.JS 3.3.0
Original install method : Elastic Cloud
Some of our errors being incorrectly grouped together. It seems the errors that get grouped together have the same error class (standard Node.JS Error
) and the same stacktrace ([1] i.e. an external service may produce multiple different errors which we then turn into an Error
class which would have the same stacktrace). The error messages are completely different. This is a large problem as we are not being correctly notified about error groups through Watcher & we have to manually go into Discover to view the error log messages to work out the correct error grouping.
Some of the errors are being grouped correctly (different stacktraces) even when the error message has an id that varies (i.e. my product had a problem (id: "f37i87"
) which is great.
My Go isn't too great but it looks like a fair amount of the grouping relies on the stacktrace and error class: https://github.com/elastic/apm-server/blob/master/model/error/event.go#L331
I can provide an example of some of the incorrectly grouped errors if required but would only be able to do this privately.
Is there a better way to handle [1] (above) to enable correct error grouping?
Many thanks, Matthew