We run cross collaboration teams, so here is a snapshot from three of our backend teams in different areas.

Team 1: We are wrapping up a migration from Ruby/Heroku to Python/AWS/AWS Lambda this year. The list of services we've got on Production that's already based on these technologies is growing fast. We're using RDS, Dynamo, and Redis, to solve our data storage and caching needs, choosing the right solution for each problem. In 2019 we've experimented with Node and Go, and believe polyglot programming can play an excellent role if you use it at the right spots. We rely on SNS and SQS for cross-service data hydration, and continue to improve our AWS infrastructure to serve our growing scalability needs.

Team 2: This year we're looking to bring full visibility into our platform by redoing our device management infrastructure, incorporating frameworks such as OSQuery and Kollide Fleet. We're also looking to improve our capabilities in gathering and presenting data on other aspects of IT such as network, security & compliance, and SaaS licensing.

Team 3: The main challenge is to build a next-generation real-time support solution. We're looking at serving our customers through multiple chat technologies, as well as other modes of communication. When we provide IT support to our customers, there are several needs and goals. We need to understand what the problem is. We need to assign the right level of technician. And we need to provide automated solutions to common requests. The focus, consequently, is on two fronts: one is building a communication-agnostic platform, through solutions such as a chat middleware (Python, AWS Lambda) and some API integration such as with Slack API and MS Teams. On the issue-resolution front, we're focusing on a series of services that can replace manual work with a series of integrations both with device management software (Jamf, Kaseya) and with various SaaS APIs.

There’s no shortage of work to be done within SRE! On the tech front, this year we are looking to migrate from Ruby on Rails (Heroku) to Python on Lambda (AWS), improve the architecture of our serverless cloud infrastructure, and improve the reliability and scaling capacity of our MDM-as-a-service offering (which is expected to grow by about 10x over the next two years). To do this, we are employing infrastructure-as-code tools such as Terraform and Serverless Framework to manage Heroku and AWS resources (Lambda, RDS, SQS, and EC2), and developing new technical solutions for de-coupling infrastructure and resource management from the development chain.

Perhaps even more important than the technical stuff, a continuing focus for our team will be practicing SRE at the cultural level. While tooling is certainly important, what separates SRE from “DevOps” is a certain mindset that everything can (and should) be measured, and those measurements should be used to give continuous feedback on service reliability to everyone in the company. With our brand-new SRE team, we are helping to encourage this mindset within engineering at every level of our stack and the infrastructure it runs on. This will require significant effort in building and improving observability, monitoring, alerting, and incident management processes, along with all the technical bells and whistles that come along with it.

The Data Team is responsible for laying the foundation for a data-driven product culture at Electric. In 2020, a big emphasis for us is partnering with the product and design teams to improve the way we capture and structure data in our product, so that we're set up to pursue machine learning opportunities in the future.

We also play a role in each feature development cycle, ensuring clearly-defined metrics are available for everything we build, and that the insights we need to make decisions are at our fingertips.

To support all of the above, we rely on a modern data stack, including Redshift, dbt, Airflow, Segment, Stitch Data, and Periscope. Our most pressing data engineering challenges involve making sure we can deliver timely and accurate insights despite the complexities of our distributed microservice architecture.