Maintainer of octoDNS and requests-futures. 20+ years of open source and professional experience across a wide array of topics.

One area of work the past few years was traffic engineering. octoDNS started as part of an effort to enable split authority DNS. The goal being to increase reliability the DNS and gracefully handle provider failures along with making DNS changes a safe and reliably process such that anyone at GitHub could drive.

I've done a lot of work scaling and operating HAProxy taking GitHub from a single North American POP to a global edge network bringing substantive performance and reliability improvements to users. We also managed the internal load balancing and between the two were running clusters that handled hundreds of thousands of requests per second.

One unifying themes through all of my work has been a focus on observability. Having a good picture of the state of a system is essential to keeping it healthy and figuring out what's up when it isn't. This is especially true when faced with substantial DDoS attacks, whether volumetric or targeting the application layer.

As part of that focus on observability I've recently been exploring vector and contributed a statsd style metrics aggregation transform to that project. That PR came about as part of my exploration work on ross/haproxy-logging-and-metrics, which I hope to turn into a canonical example of HAProxy logging best practices. I also have plans to build a broader HAProxy configuration & operation best practices example in ross/haproxied, but as of yet it's very early days.

I'm always happy to chat about all things observability, DNS, HAProxy, or most anything tech. I'm not currently looking for employment opportunities, but would be open to chatting about consulting and/or contracting on projects that intersect my interests. My contact info can be found on my profile. For more information on my work & project history see LinkedIn.