Webinars Logs Debugging

Surface and Confirm Buggy Patterns in Your Logs Without Slow Search

Summary:


Debugging with logs in distributed systems can be a pain. It’s tough to search raw data looking for a pattern, relating potential causes with other logs, and checking trace and metrics data for more confirmation. Is finding one pattern enough? What if there are other problems? Who knows how many colliding factors are relevant? At Honeycomb, we’re flipping the script on the log search problem. Hear our resident experts, (former Splunk Ninja) Michael Wilde and Andy Dufour, discuss how Honeycomb customers have technically evolved their log analysis process to achieve fast pattern detection, skipping the search grep/search loop entirely. Learn how you can begin doing the same, no matter how mature your log instrumentation is today.

Topics include:

- How to technically prep your log structure for faster analysis
- Practical paths to send useful log data to Honeycomb
- What attributes to analyze in your log data at different levels of instrumentation
- Moving up the logging maturity curve with achievable steps, even if you’re at unstructured, unpatterned raw strings today
- How engineers at Vanguard and SumUp went through this process to transform the way they work

Learn how you can implement a fast and accurate debugging process that tracks complex application requests while taking full advantage of the rich context within your logs.