Some time ago I wrote about inadequacy of our error reporting infrastructure. In the comments, Igor advised us to use log collectors such Splunk or Logstash. We have seriously considered this option, but we find Splunk too expensive and Logstash incompatible with our software stack. After some more research we ran into Sentry that looks like a solution to our error reporting problem. Sentry is an open-source project that originates at Disqus (We share our NYC office with a part of Disqus team!) and concerns itself with realtime event logging and aggregation. Sentry is implemented using Python, Django, and Celery, which makes it perfect fit to our environment. Additionally, the guys at Disqus spun a new company providing hosted Sentry service which gives quite some confidence that Sentry will continue its rapid development. Adding support for Sentry in your application is quite trivial, since the only thing that you need to do is to add a logging handler that sends log statements above a specified level to the Sentry server. All other configuration and management is done thru Sentry web application and is independent of your code. We've only installed Sentry a few days ago, so this post is not a review based on actual experience. On the contrary, it is more a call for advice if any of you have a first hand experience with Sentry that you can share with us.