As soon as it is clear that the R code has failed to run the datanode UI should update to reflect this (i.e. execution failed). Instead what I currently see is that analyses will fail which is clear from the docker logs of the datanode container but the UI still says "executing" for quite a long time.