The Blog: October 2009
Spelunking into your logs with Splunk
At Agora we’re experimenting with Splunk as a error collection and reporting tool. The idea is that all our services will spit out error information to syslog which will be picked up and indexed by Splunk. Splunk will be configured to display info about recent errors, email summaries of those errors and generally be the starting point for discovering problems with our systems. In our previous infrastructure error collection and reporting was handled on a per-project basis. As we’re now moving to a more service oriented approach that involves a small number of generic but highly configurable services deployed over a large number nodes we needed something different. Splunk seems ideal for this. It accepts a number of different data sources ranging from syslog to raw log files to inputs from random UDP and TCP ports so we can easily integrate with systems we’ve written and, more importantly, those we haven’t. It also has the summary and reporting capabilities we were looking for so we can chart errors over time periods, dig for trends and see which services are acting up.
Integrating Lighthouse and Instant Messenger in Ruby
The Introduction to the Lighthouse API mentions something you might do with the Lighthouse API is “Accessing and creating tickets through your Instant Messenger client.”. How easy is it to do this? Surprisingly easy.
Mocking HTTP request and response with Python and Mox
I recently started integrating with some web services in Python and wanted to be able to test all the requests and responses with valid and invalid data. However, I need to be able to do this without actually hitting the web service when testing. And of course I don’t want to change my web service code to conditionally do things if running in a test environment. Enter mocking and Mox.