How the cloud team at 8x8 gained visibility into service maturity with Cortex

8x8 is a cloud-based platform on a mission to transform the way businesses communicate. Its secure and reliable cloud platform integrates voice, video, chat, contact center, and enterprise-class API solutions to connect people around the world.

What was the problem?

Over time, the cross-functional cloud observability and reliability team at 8x8 struggled to keep track of all the data it received from the DevOps, ops, support, and engineering teams. Without a structured data store, the team couldn’t identify where components belonged and who was responsible for those components.

This information existed, but it existed in a fractured state, spread across wikis and Git pages. There was no consistency in cataloging efforts, making it more difficult to draw conclusions from the data the team did have. For the most part, individual team members stored crucial information in their heads.

How was this solved before?

8x8 implemented a few different solutions before using Cortex, but those solutions varied by VP and organization. One VP built a service catalog in a wiki that contained internal maturity and scalability ratings, operational concerns, links, and other metadata — but the catalog was only for this VP’s organization, and it relied on self-reporting.

For the most part, operational teams at 8x8 built spreadsheets to track ownership and service metadata. Inconsistency across spreadsheets and incomplete data made it difficult to gain traction on solving problems. At this point in time, the solution depended on who was asking the question.

How did scorecarding help with service maturity at 8x8?

Before the cloud team could address operational maturity, they had to confront the fact that there was decades-old code living alongside brand-new code. Their first steps were focused on development maturity — understanding where the code lives, who owns it, what goes into it, and how it’s all related.

The cloud team started scorecarding based on development maturity, making sure all the basics were there. This meshed well with the process of onboarding with Cortex. If a service wasn’t in Cortex, then it couldn’t be measured, so it received a zero grade for development maturity by default. 8x8 found that integrating with Git was crucial at this stage, especially — if they didn’t know where the code was, a service was definitely in trouble.

It took some time for everyone to come to a consensus around the basic maturity standards, which is only natural when you first bring together a bunch of teams with unique processes and perspectives. After coming to an agreement around best practices and quality standards, it has become second nature for team members to drop new services into Cortex.

How has the catalog sparked new initiatives at 8x8?

“Just knowing what’s there and being able to start asking questions about it — I think Cortex has paid for itself in terms of having easy access to that data, and being able to store and retrieve the data, and link things together.”

8x8 first approached Cortex with a simple goal: the cloud team wanted to point to a service’s maturity score and provide straightforward steps for getting any service in line. However, the real value of Cortex for this time was actually in the catalog.

They found that the catalog enabled team members to properly classify services and accurately identify scopes of concern for each service for the first time. The visibility the catalog provided prompted the team to develop new initiatives. For example, they could now identify all the tier-1 services with a clearly defined, critical SLA, which enabled them to ask questions about those services. Before using Cortex’s catalog, they didn’t know what questions to ask.

Having this data at their fingertips led team members to consider best practices for tracking metrics for engineering management, including development cycle, release cycle, MTTR, and change failure rates. The catalog that prompted team members to ask questions about testing for adherence to these metrics in production. As a result, 8x8 now has a huge game day initiative designed to make sure services are ready to go.

Did you learn anything unexpected along the way?

“Being able to see the data and work with it from different dimensions really opened up a whole new world of possibilities.”

For the first time, management could see how ownership worked in practice. Cortex revealed that one owner had more than 50 services, but only one was a tier-1 service; another owner had only five services, but all of them were tier-1 services. This allowed them to consider how they could make ownership more efficient, and whether teams had the resources and developers needed to manage their resources.

What’s your favorite feature of Cortex?

“Cortex is one of those things where it’s like, you’d never know you needed it until you had it. And as soon as you have the service catalog, you’re like, ‘How did we live without this thing?'"

Because so much of the cloud team’s work involves automation, Cortex’s API has become a crowd favorite. Being able to structure the API in order to pull out precise data has saved a lot of time and energy. The cloud team often uses the API to build reports detailing ownership and other customized metadata so they can gain even more precise insights into their services.

What's the next stage in your service maturity journey with Cortex?

“When VPs ask a question, they can trust the reporting from Cortex to help them find their answer. They know the information is consistent, and you can’t put a price on that level of consistency. It’s really important…at least we know this information is real, and it came from the system, rather than just somebody pulling it out of their head.”

By working with Cortex, 8x8 was able to really trust data about their services for the first time. VPs used to dread spreadsheets and reports, but with the data integrity and consistency that Cortex offers, they had the ability to make use of this information. Rather than survey dozens of different people, they can now get the answers to their questions quickly and directly — and they can trust those answers are accurate.

8x8 has enjoyed the personal relationships they’ve been able to build with members of the Cortex team. They’ve been able to help Cortex build out new features and get feedback more quickly. Cortex is excited to continue working with the team at 8x8 to make our platform even better for all of our users. If you’d like to see a demo of Cortex and how it can help your teams with service quality, book a demo today!

Matthew Gates
Matthew Gates - Principal Engineer