John Nunemaker

Rails: Following and Notifications

A month or so ago we released following and email notifications for Speaker Deck. I thought I'd share how we whipped this together and why we did it the way we did. The end result is a stable notification system that gets the job done and will be easy to grow as Speaker Deck continues to grow.

rails

Rails: Easy Sitemaps

I've been doing some reading and learning about SEO of late. An easy thing to do is to create a sitemap and submit it to google's search console. Here is how to do that with Rails.

TIL

Rails: insert_all and upsert_all

TIL something new about Rails. You can now use insert_all and upsert_all for efficient writes. But that's not all. Read on for what "just worked" and what didn't.

imgproxy

Imgproxy is Amazing

One of the cruftiest pieces of the Box Out rails app was file previews (for customer uploads and created graphics). Background jobs, backups, errors, and waiting are all bemoaned and later stupified (sorry I just watched Harry Potter) in this post. The Needles: So

flipper

More Descriptive Flipping

The flipper ui just got a fresh coat of paint and I am here to tell you about it. From adding descriptions to your features to bringing the OSS UI more in line with flippercloud.io, flipping features just got even better.

performance

Never Query the Same Thing More Than Once

In a recent post, I talked about how I used Postgres DISTINCT ON to remove a pesky N+1 query. That was actually one of the final touches in immproving the Graphics#index action for Box Out. Prior to that, I whipped together a

performance

Limit Everything: Timeouts for Shell Commands in Ruby

I have talked about limiting everything before, but I love examples, so I thought I would start bringing some out of the woodwork. How many times have you whipped out backticks or one of the million other ways you can shell out in Ruby?

performance

Postgres Distinct On

By day, I am a mild mannered owner/programmer at Box Out Sports. By no means do you need to understand all that we do there or how we do it, but recently I fixed a tricky N+1 query and thought I should

flipper

Flipper Cloud: Environments

I will be honest with you. I really miss writing. There is something about organizing the ideas and work swirling around in your head into a blog post that makes the abstract concrete and the chaotic organized. I mean shoot, I remember the days

flipper

Flippin' and Rollin'

Two months ago I wrote up a doozy of a post about flippin’ feature at runtime. Since then, the contributions to Flipper from myself and the community have continued. Easy migration from Rollout, bug fixes for several adapters (ActiveRecord, Sequel and RedisCache) and new

flipper

FlipperCloud.io: Why?

I have moleskines full of ideas. When I worked at Notre Dame (circa 2006), making products was all the rage. I spent all my free time thinking up and working on new apps. They never went anywhere, but were great fun and, more importantly,

flipper

Flippin' Features at Runtime

Since it has been nearly a year since I’ve written about Flipper (a ruby gem for turning parts of your application on or off at runtime), I thought I’d share some of what I (and other contributors) have been up to. From

performance

Resilience in Ruby: Handling Failure

For the second half of 2014 and 2015, most of my time at GitHub was spent working on moving the notifications feature off the primary MySQL cluster on to a new MySQL cluster and making the feature resilient to failure of that cluster. When

performance

Fewer and Faster

Sometimes I feel developers think that performance is a dark art. It is not. In my experience, well performing systems come down to this: fewer and faster. If you are doing something a lot, do it fewer times. If you are doing something that

randoms

Tuesday Randoms: 2016-04-19

The Error Model. I’ve been doing a lot of resiliency related things lately and that means dealing with errors. This post is one of the longest (and also best) summary of error handling on the planet. Write code that is easy to delete,

randoms

Tuesday Randoms: 2016-01-19

Fail at Scale Great post about failure at Facebook. My favorite parts were controlled delay and adaptive LIFO. These two data points seem to suggest that when Facebook employees are not actively making changes to infrastructure because they are busy with other things (weekends,

performance

Anti-Decay Programming

de•cay noun - the state or process of rotting or decomposition. Let us kick off the new year with a lovely post on decay. All systems decay. Even when you continually work on improving the system, it decays. That said, in my experience,

performance

Database Performance Simplified

I have been fortunate enough to work on a few high throughput applications (Words with Friends, GitHub.com, Gaug.es, etc.; hundreds to thousands of requests per second) in my life where a database was the bottleneck. I am far from an expert on

performance

All Requests Are Not Equal

The norm for rate limiting these days seems to be based on the number of requests a consumer can make in a given time window. The problem with this is that all requests are not equal. Some requests are more expensive than other requests

randoms

Saturday Randoms 2015-12-07

Previously. QuotesThe true purpose of testing, just like the true purpose of design, is to reduce costs.Sandi Metz in Practical Object-Oriented Design in Ruby. MotivationalWant to Create Things That Matter? Be Lazy. Love the splitting of work into two categories: deep and shallow.

randoms

Monday Randoms: 2015-11-02

Previously. ResiliencyResiliency checklist A fantastic checklist broken up into 4 categories: fault tolerance, scalability, operability and efficiency. What happens when a dependency starts failing? What if it begins failing slowly? How can the system shed load? What is the chief metric with which the

randoms

Wednesday Randoms: 2015-09-30

Previously. DataBuilding DistributedLog: Twitter’s high-performance replicated log service How Twitter build a distributed log on top of Apache BookKeeper. Great post on the features they required and the results. In this section we describe DistributedLog, a serving layer that we built on top

programming

When and Why Your Code Starts to Smell Bad

I enjoy reading papers and try to on a regular basis. I actually read them old school style. I print them and then use a pen to underline the quoteables that I liked. Sometimes I annotate them with thoughts, and rarely, I have even

randoms

Saturday Randoms: 2015-09-19

I read a lot. I curate what I read into likes in instapaper and bookmarks in pinboard. It hit me the other day, as I was reading one of the many email newsletters I subscribe to for dev tips, that there is value in

flipper

Flipper: Insanely Easy Feature Flipping

Nearly three years ago, I started work on Flipper. Even though there were other feature flipping libraries out there at the time, most notably rollout, I decided to whip up my own. Repeating others is, after all, one of the better ways to level up your game.

Subscribe to John Nunemaker