“We’re now deploying 3 to 4 times more a day”
“Deployments have become something we don’t need to worry about. They just run. This means we’re able to deploy over 100 times per day. ” - Aaron Cline, Sr Director, Site Reliability
Sojern is a marketing platform for hospitality that helps users find, attract, convert, and engage customers throughout the travel buying journey. Its platform delivers marketing messages directly to customers. The platform maximizes growth and revenue through unrivaled traveler insight, intelligent audiences, multichannel activation and optimization, and a connected hotel guest experience—all in one place.
The technological backdrop
Sojern has always deployed to Kubernetes, but before Codefresh, its teams used an open-source tool to manage their pipelines and deployments. They’d built a mono repo and monolith pipeline that everyone used for deployments. When they only had 2 teams, it was manageable, but as the company scaled, it started to run into issues. Teams wanted to move to Argo CD, but also wanted a tool to help standardize deployments while using industry best practices.
Challenges
Sojern wanted to modernize its approach to scale deployments
Before using Codefresh, the system Sojern inherited was built on a lot of custom code. Teams often ran into issues with infrastructure maintenance. The site reliability engineering (SRE) team that supports the developers often had to jump in and support failed deployments, which became difficult to manage.
Sojern was scaling and bringing on new teams, all deploying the same monolith pipeline and mono repo. They were often held up by manual processes that were in place to avoid failed deployments. They needed a tool to help them scale effectively, as the open-source tool they used wasn’t keeping up with their fast pace.
The team wanted to migrate to Argo CD and split out their pipelines across teams, while implementing industry best practices into their processes. They wanted a tool to simplify the migration and automate their manual deployments. They also needed the tool to support Gerrit, which they used for web-based team code collaboration.
Solution
Sojern uses Codefresh to deploy over 100 times a day using industry best practices
After evaluating other tools, Sojern chose Codefresh to migrate to Argo CD. They used this to modernize their Continuous Delivery (CD) without needing a lot of in-house experience. Codefresh made the migration process easy, so they could get the benefits of Argo CD while hosting and supporting it.
Using Codefresh, teams have standardized on their deployment processes, so everything is consistent and meets industry standards. Their deployment processes are automated, they’ve removed delays, and split out their large monolith pipeline.
With Codefresh, Sojern has increased its number of teams as manual processes no longer bog them down. The organization has grown from 2 teams to 10, with an additional 20 developers.
Sojern has also implemented GitOps using Codefresh. Developers no longer manually set up manifests, as much of the key information gets automatically filled in for them. With monitoring in place, developers can quickly see all the deployment information they need. If there’s a failure, they get a notification on Slack and can quickly troubleshoot. Team members with permissions can also promptly restart or stop pods when needed.
“Using Codefresh means we need to worry less about the deployment process and how it’s maintained. It’s stopped the manual process hurdles for us.”
Nishant Yedla, Staff Site Reliability Engineer
Value
More frequent, automated deployments
By automating deployments with Codefresh, Sojern now deploys 3 to 4 times more each day than it used to. Teams also regularly deploy to production over 100 times a day. This wasn’t possible with their old system and their manual processes, as they could only deploy one pipeline at a time. Now, deployments are automated, easy, and fast.
“We used to be able to only deploy one pipeline at a time, and for each deployment, we had to manually focus on health, rather than letting our tool handle that. Now, it’s become easy. It just runs.”
Aaron Cline, Sr Director, Site Reliability
Simpler management of security and auditing requirements
Sojern has several security policies it must adhere to for its SOC 2 compliance. With Codefresh, it can ensure these are met. Codefresh lets them control the teams that can deploy, run automated testing, and fail a deployment if tests fail. This has made maintaining security requirements quicker and easier.
Sojern also values the audit trail of deployments. Its previous tool made keeping track of job histories difficult. With Codefresh, teams can look back and see all the information they need about deployments.
“Being able to look back at every deployment and see all the historic data is very valuable to us. It’s not something we had with our previous tool, but it’s easy with Codefresh.”
Aaron Cline, Sr Director, Site Reliability
Improved developer experience
Before Codefresh, Sojern’s deployment processes were manual and time-intensive. Using Codefresh, teams have implemented industry best practices and automated their deployments. This has reduced the manual burden so they can focus on shipping more features.
Teams also had to manage scheduled maintenance windows with their previous tool, which created 20–30% downtime. With Codefresh, their downtime is less than 5%, so developers can deploy more frequently.
“Supporting our developers across deployments wasn’t the biggest issue; it was more when something went wrong. Now that we’re using Codefresh, we’re following industry standards and our deployments are faster and simpler for teams.”
Aaron Cline, Sr Director, Site Reliability
Reduced risk of failures in deployments
Before Codefresh, failed deployments were common, as Sojern’s system was designed for smaller teams. As its teams grew, the solution didn’t scale, and deployments often failed. Now, failed deployments are rare. If deployments do fail, they can perform a rollback in under 10 minutes to avoid downtime.
“Our previous solution just didn’t scale, and many of our teams ran into each other. It was just a mess. Codefresh has helped make many of those problems disappear, all while we’ve scaled.”
Nishant Yedla, Staff Site Reliability Engineer