Plans for a logback appender? about raygun4java

[fundead]

Hi, sorry about the delay in getting back to you on this - it's planned for the near very near future due to it being such a common use case. We're uncertain as to whether it will end up in this repo as a Maven package or not, as the scope of frameworks Raygun4Java supports is starting to grow, but if you would desire it from this Maven package let me know.

Regards,

Callum Gavin
Mindscape Limited

[gkopff]

I'm not fussed where the appender resides - I assume it will be published to a public maven repo?

Do you plan to support both exceptions as well as log messages? I think capturing some information around the message call-site so that the same log message can be coalesced into a single item (with the details about 'reported n times') is a useful feature. As I said originally, it's not only exceptions that we want to catch and handle like this - it's any unexpected event.

[fundead]

Unfortunately as Raygun currently stands the architecture has first-class support for the capture and presentation of exceptions - we are definitely aware of the desire for a generalized logging framework but this is somewhat outside our current scope in the near term.

Logging presents a different challenge due to the limited data present; our automatic grouping relies on the stack trace and related data of exceptions to group without using the message. Using the message does not provide a good enough experience as instance-specific data is often present (user or context information), therefore the service is designed around automatic grouping and wholesale notifications/resolving of exceptions within those error groups as they have the necessary data available.

As I've mentioned generalized logging is desirable but the service does not currently have back-end or UI support for this, but we have it in mind for an additional feature in the future. There's a feature request here that you could add your voice to: http://raygun.io/thinktank/suggestion/2776 . Any suggestions will be taken into consideration when we discuss the architecture and UX surrounding this feature.

[gkopff]

Logging presents a different challenge due to the limited data present; our automatic grouping relies on the stack trace and related data of exceptions to group without using the message. Using the message does not provide a good enough experience as instance-specific data is often present (user or context information), therefore the service is designed around automatic grouping and wholesale notifications/resolving of exceptions within those error groups as they have the necessary data available.

I made this work with my airbrake appender by getting the stacktrace to identify the log message callsite, and passing that information along with the message. This obviously has some overhead, but overhead I was prepared to pay since this was a serious error.

[gkopff]

I've made a raygun equivalent of my old (closed-source) airbrake appender here: https://github.com/gkopff/logback-raygun

[fundead]

Thanks for open sourcing that, I'm sure others in the community will greatly appreciate that. Will close this issue for now.