I want to deploy a squash server for us so we can track our agent (and other) exceptions. This will make debugging agent crashes much quicker (at least for me because I am bad at log parsing).
This does some cool things like:
Count exceptions (grouping them by stack trace and message)
Attempt to figure out who's responsible using git blame and send an email (if desired)
Store the state of the system when the exception occured (less useful for Smith, but good for web apps)
Discussed offline and decided this is slightly outside of the scope of smith. As such a more flexible solution allowing agents to hook into the exceptions should be implemented.
I want to deploy a squash server for us so we can track our agent (and other) exceptions. This will make debugging agent crashes much quicker (at least for me because I am bad at log parsing).
This does some cool things like:
git blame
and send an email (if desired)