30 Aug 2017
How on earth does someone accidentally delete 85% of their users’ GitHub tokens? I was suspicious that something might be wrong when I got an email from a service I run called CodeTriage, it’s a free web app to help find open source projects and issues to work on. While I get plenty of emails from my service, I don’t often get ones with the subject line “Code Triage auth failure”. Before we can understand what happened, let’s look into why this email even exists.
Keep Reading
23 Aug 2017
If you use twitter, you’ve likely seen those tweets with the little *promoted* tag on them. Twitter has always been a huge source of traffic to my tech articles, so I wondered if a promoted tweet or two would be just as successful. To find I went down the twitter advertising rabbit hole for the first time. Keep reading to find out how well my promoted tweets performed.
Keep Reading
15 Aug 2017
This is a post about how I had a bug in a background job that ended up sending huge emails to customers. Learn about how I found & debugged the issue, mitigated the problem, and then finally fixed the underlying causes.
Keep Reading
09 Aug 2017
I’ve not seriously touched frontend code, in years. Frankly, it scares me. To that end “front end devs are not real programmers” is totally BS. I want to talk about some of the recent changes in tooling and APIs that are available so that front end development might not suck as much as it used to. You will not learn to be a CSS or JS guru with this post. If you’ve written much front end code, this will be mostly full of face-palm level obvious statements. Therefore, feel free to read for the laughs.
Keep Reading
01 Aug 2017
WEBrick is the “slowest” webserver in Ruby, how could it possibly be webscale? To answer this question and explore Is Ruby Too Slow For Web-Scale?, we will compare WEBrick to a real piece of “webscale” tech: NGINX.
Keep Reading
26 Jul 2017
Ever wonder if you should go back to school to get a master’s degree? Right now I’m in my second semester of Georgia Tech’s Online Master of Science in Computer Science (OMSCS). I’ve had a few people ask about my experiences so I figured it was worth my time to write them down. In this post I’ll be going back and forth, question and answer style to share my thoughts on the program.
Keep Reading
18 Jul 2017
Database load can be a silent performance killer. I’ve been optimizing the query performance of a web app I run designed to get people involved in open source, but was seeing random spikes of query times to 15 seconds or more. While I had been seeing this behavior for some time, I only recently began tuning my database queries. You can read about my efforts to First I sped up my home page with some indexes (and Rack Mini Profiler). Then I tracked down and killed some expensive queries. After these major improvements the average response time was around 50ms and my perc95 was under 1 second. Yet, I had this annoying issue where in a 24 hour period, my perc95 response times would shoot up to maybe 15 seconds or 30 seconds and start timing out for a short period of time. This post is about me finding and fixing that issue which resulted in a net 80% decrease in my database load.
Keep Reading
11 Jul 2017
I recently demonstrated how you can use Rack Mini Profiler to find and fix slow queries. It’s a valuable tool for well-trafficked pages, but sometimes the slowdown is happening on a page you don’t visit often, or in a worker task that isn’t visible via Rack Mini Profiler. How can you find and fix those slow queries?
Keep Reading
28 Jun 2017
Update: I did mention that lock free data structures are really hard to write, it looks like there might be some issues that haven’t been addressed in the implementation of this LF Queue that we’re referencing. The rest of the analysis is still valid and hopefully useful to you, just know there’s actually more that needs to be done, don’t try to use that code for a mission critical application out of the box.
Keep Reading
22 Jun 2017
When I see a query in my logs without either a limit
or a count
clause, alarm bells go off because it is likely a hotspot. A pagination query has alimit
so it usually flies under my radar:
Keep Reading