Converting error analytics into tickets at-scale?

[Noble_Draconian]

Hey everyone! I have a bit of a tricky issue that I haven’t been able to figure out the solution for.
I was wondering what workflows you folks used to convert errors in your analytics into tickets, at-scale?

Currently in my game, whenever an error occurs, that error is logged to my analytics (I use playfab). This works fine for error tracking, but it has one issue : It isn’t that scalable. After a week or so, the errors for the version in production become quite plentiful:

Because I use playfab, there’s no way to really mark errors that have had tickets made for them as “read”. As such, I end up having to scroll through the many errors each day, manually looking for any NEW errors that haven’t had tickets made for them yet. This is obviously a very error-prone and time-consuming process. New errors can pop up above old ones (since we sort by error frequency in descending order) if they occur more frequently.

I was wondering what you folks do to convert errors into tickets at-scale? How do you automate this? Do you copy the raw json data from the query & dump it into some sort of a tool? Do you have a dedicated person for managing tickets? Something else?

Thanks in advance for your responses, looking forward to them!

[grilme99]

I’m a little late to reply but PlayFab isn’t the solution for error tracking here.

You have a couple of options. The free one is to keep a datastore of all errors that have occurred thus far and create GitHub issues (or whatever else) from in-game via the relevant HTTP APIs when you get a new error.

Ideally, you would use a platform made for error tracking (such as Sentry). This can be costly but is well worth it as a project grows. Sentry provides all sorts of integrations for you to make use of. This definitely isn’t an option for a lot of games due to the cost, but if you can use it then I highly recommend it. Keep in mind you can also self-host Sentry on a VPS for a lot cheaper than their cloud solution.