Lunch Talk Series: Cutting latency in half: What actually worked—and what didn’t

This iteration of the Lunch Talk Series will be covering Timothy McCurrach’s latency-reducing DjangoCon US talk, “Cutting latency in half: What actually worked—and what didn’t”.

You should watch this if:

• You’re interested in improving performance proactively.
• You’re interested in making more reliable and robust web apps.
• You’re looking to begin profiling proactively rather than reactively.

Timothy starts the presentation off by discussing reactive profiling versus proactive. His own story resonates with mine quite a bit, so I found myself quickly wanting to adopt his proactive processes.

I 100% love his mindset on how he approaches making robust applications. It’s well-reasoned, practical, and honest. He explains that the tools aren’t all that special, but your knowledge of how to use them is. Timothy encourages others to explore to to make your tools work better for you. Another recommendation was to build a toolkit for yourself, he shouts out the recent addition to Django to allow people to auto-import helper functions which he makes use of for debugging performance problems.

Another fascinating observation from Timothy was that you have likely fixed all the things that you’re aware of. The rest of the problems are the things you aren’t familiar with. This is why you haven’t dealt with them. I’m not sure I had thought of that scenario in that way before, so I found it quite insightful.

For me, my takeaway is that I need to implement solutions that highlight performance across the site more systematically. Being aware of how things operate and making it easy to access will improve my decision-making and, hopefully, the performance of my applications.