Kafka

Lumosity is home to the world's largest cognitive training database, a responsibility we take seriously. For most of the company's history, our analysis of user behavior and training data has been powered by an event stream--first a simple Node.js pub/sub app, then a heavyweight Ruby app with stronger durability. Both supported decent throughput and latency, but they lacked some major features supported by existing open-source alternatives: replaying existing messages (also lacking in most message queue-based solutions), scaling out many different readers for the same stream, the ability to leverage existing solutions for reading and writing, and possibly most importantly: the ability to hire someone externally who already had expertise.

We ultimately migrated to Kafka in early- to mid-2016, citing both industry trends in companies we'd talked to with similar durability and throughput needs, the extremely strong documentation and community. We pored over Kyle Kingsbury's Jepsen post (https://aphyr.com/posts/293-jepsen-Kafka), as well as Jay Kreps' follow-up (http://blog.empathybox.com/post/62279088548/a-few-notes-on-kafka-and-jepsen), talked at length with Confluent folks and community members, and still wound up running parallel systems for quite a long time, but ultimately, we've been very, very happy. Understanding the internals and proper levers takes some commitment, but it's taken very little maintenance once configured. Since then, the Confluent Platform community has grown and grown; we've gone from doing most development using custom Scala consumers and producers to being 60/40 Kafka Streams/Connects.

We originally looked into Apache Storm / Heron , and we'd moved on from Redis pub/sub. Heron looks great, but we already had a programming model across services that was more akin to consuming a message consumers than required a topology of bolts, etc. Heron also had just come out while we were starting to migrate things, and the community momentum and direction of Kafka felt more substantial than the older Storm. If we were to start the process over again today, we might check out Apache Pulsar , although the ecosystem is much younger.

To find out more, read our 2017 engineering blog post about the migration!

I started using .NET in the early 2000s. Ever since version .NET 3.5 (and even .NET 2.0 if we take a proper generics implementation into account), C# was dominating in the feature battle against its rival, yet wasn't advancing significantly in the product coverage due to its platform dependency.

Thus I was very excited to hear the news about plans to develop an open-sourced cross-platform .NET Core framework. We started using .NET Core in production from version 1.1, and a global decision to migrate the entire solution to .NET Core was made with the release of .NET Core 2.0. Now we have more than 100 .NET Core (micro)services running on Linux containers inside Kubernetes, using Kafka for reactive communications and a number of open-source relational and NoSQL storage engines.

When I joined NYT there was already broad dissatisfaction with the LAMP (AngularJS MySQL PHP) Stack and the front end framework, in particular. So, I wasn't passing judgment on it. I mean, LAMP's fine, you can do good work in LAMP. It's a little dated at this point, but it's not ... I didn't want to rip it out for its own sake, but everyone else was like, "We don't like this, it's really inflexible." And I remember from being outside the company when that was called MIT FIVE when it had launched. And been observing it from the outside, and I was like, you guys took so long to do that and you did it so carefully, and yet you're not happy with your decisions. Why is that? That was more the impetus. If we're going to do this again, how are we going to do it in a way that we're gonna get a better result?

So we're moving quickly away from LAMP, I would say. So, right now, the new front end is React based and using Apollo. And we've been in a long, protracted, gradual rollout of the core experiences.

React is now talking to GraphQL as a primary API. There's a Node.js back end, to the front end, which is mainly for server-side rendering, as well.

Behind there, the main repository for the GraphQL server is a big table repository, that we call Bodega because it's a convenience store. And that reads off of a Kafka pipeline.

At Heap, we searched for an existing tool that would allow us to express the full range of analyses we needed, index the event definitions that made up the analyses, and was a mature, natively distributed system.

After coming up empty on this search, we decided to compromise on the “maturity” requirement and build our own distributed system around Citus and sharded PostgreSQL. It was at this point that we also introduced Kafka as a queueing layer between the Node.js application servers and Postgres.

If we could go back in time, we probably would have started using Kafka on day one. One of the biggest benefits in adopting Kafka has been the peace of mind that it brings. In an analytics infrastructure, it’s often possible to make data ingestion idempotent.

In Heap’s case, that means that, if anything downstream from Kafka goes down, we won’t lose any data – it’s just going to take a bit longer to get to its destination. We also learned that you want the path between data hitting your servers and your initial persistence layer (in this case, Kafka) to be as short and simple as possible, since that is the surface area where a failure means you can lose customer data. We learned that it’s a very good fit for an analytics tool, since you can handle a huge number of incoming writes with relatively low latency. Kafka also gives you the ability to “replay” the data flow: it’s like a commit log for your whole infrastructure.

#MessageQueue #Databases #FrameworksFullStack