Exception Tracking (Sentry)

[Jesse Kretschmer]

Hey Guys,

I was watching that old thread about the python logging module be resurrected, and I wanted to point out something that I found which is awesome! If you are feeling adventurous, you should have a look at this thing called: Sentry (https://pypi.python.org/pypi/sentry). 

What is it?

Sentry is just a messaging service. It's meant to aggregate a large number of messages produced by code and show them in a simple feed.

Why does it matter?

It is a really smart solution to a question I had never really asked; "How do I capture bugs before users report them?" There is even a paid service, but luckily the underlying code is all open source. It's a pretty simple Django service that's supported by Postgres. I've got a background in web-dev so I was comfortable getting the service standing on a VM provided by my system operations team. The docs are pretty good, so I'm sure you all can get running without much effort.

How does it work?

As I mentioned Sentry itself is just a message service. You just need to send it good data. The same paid service also release their client Raven (https://github.com/getsentry/raven-python). It has a built in exception capture system that will step through the exception and dump the locals() per frame (Really awesome for debugging!) AND it's threaded, so it should not interfere with the user interface.

I already had a logger in all my studio code and I was already using an exception hook to capture out uncaught exceptions in a log file. The Sentry system is decidedly better.

Here's a bare-bones example of how you might implement exception tracking:

import sys

try:

    import raven

    RAVEN_CLIENT = raven.Client('http://<key>@<server>:9000/<projectId>')

except:

    RAVEN_CLIENT = None

def exceptionHandler(excType, excValue, traceback):

    if RAVEN_CLIENT!=None:

        RAVEN_CLIENT.captureException(exc_info=(excType, excValue, traceback))

sys.excepthook = exceptionHandler 

But this is a lot of data! Don't worry, Sentry will handle data expiration in a sane way. Repeat messages are dumped. I've only chewed up a few hundred megs for my studio with roughly 200 users creating tracebacks.

You can even start reporting other messages or mix in some tagging to track your projects. The tool has a rich set of features that I'm still learning.

So have a look. I hope it makes life better for someone out there. 

Cheers,

Jesse

[Marcus Ottosson]

I must admit I was a little put-off by the sales-pitch-style introduction, but I eventually clicked the link at the end and it actually looks quite useful. Thanks for the tip.

--
You received this message because you are subscribed to the Google Groups "Python Programming for Autodesk Maya" group.
To unsubscribe from this group and stop receiving emails from it, send an email to python_inside_m...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/python_inside_maya/CANESWi2ny_yAG3%2BaoEOh00jv-K7q6ANjPst%3DXwnR%3DSXeN9RVCw%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.

--

Marcus Ottosson
konstr...@gmail.com

[Jesse Kretschmer]

I apologize for being sale-sy. I inadvertently drew from the pitch I made internally at my company. 

To be clear, I am not affiliated with the project in any way. I'm a few weeks in to using it, and I wanted to share.