Better insights into error types in pack stats

Making Packs
1

As the maker for a pack I would like to be able to gain insight on the types of errors encountered by the pack users.

Currently the pack stats shows a bar graph, but this does not really provide enough information to be useful. For example, a single user of the pack could be calling the pack an enormous amount of times (which can happen as the mapping between user interactions and # of pack invocations is not 1:1) with some invalid data, causing a huge amount of errors, and causing the graph to spike. However this is the expected behaviour, as these errors are normal.

I think at the minimum this would involve either showing different graphs or data points for thrown UserVisibleErrors vs unexpected errors that have bubbled up. Even better would be tracking all the types of unexpected errors to differentiate them. And another avenue would be to be able to pass a token or number (would need to be enforced to be some static value for privacy and security) to the UserVisibleError constructor to be able to identify the prevalence of each known error type.

5 Likes
2

Thanks so much for the feedback @loucadufault, it’s great to hear first-hand what Makers look to these graphs for and where we can improve. I think separating out expected vs unexpected errors is a great first step. Error codes would be much more useful, but we’d have to think carefully about how to secure that to prevent extracting user data. I’ll pass this feedback on to the team.

1 Like