Client story Salesloft Cuts Complexity by Migrating 1PB+ of Data to Google Cloud

Empowering sales teams with better data

Salesloft leads the way in revenue orchestration. They offer a comprehensive suite of sales AI and technology solutions that help sales teams create, convert, close, and grow more revenue. As the company grew, their technology needs changed. They had to manage systems on both AWS® and Google Cloud. This split made it hard for the site reliability engineering team to operate efficiently. Salesloft decided to move everything to one cloud to stay ahead in the fast-moving tech world.

The challenge: Managing two clouds and technical debt

Managing two different clouds created a huge burden for the Salesloft team, which Phillip Burgher, senior manager, site reliability engineering at Salesloft, calls cognitive load. His site reliability engineers (SREs) and database engineers (DBEs) struggled to keep up. The team used Infrastructure as Code to build their systems, but because they worked in two clouds, they had to learn and use two different sets of tools.

“Having to learn, understand, and implement twice the platforms because we were in two different clouds was challenging,” says Burgher. “Building a storage bucket in AWS differs from doing it in Google Cloud. These small differences added up. They increased the time it took to roll out the basic parts needed for new features.”

This delay hit the feature delivery teams hard, slowing them down. “My site reliability team talks with the software delivery teams about how to build things, and using two clouds made those talks take longer,” says Burgher. “This meant customers had to wait longer to see new value from their Salesloft solutions.”

Worrying about a massive move

Beyond daily speed, Burgher worried about such a massive move. Salesloft had 120 applications and more than 1 PB of data to move, including 200 TB across multiple database clusters. They also had a huge volume of data in S3 buckets. All this raised concerns about data integrity. “Data integrity was definitely anxiety-producing because we had to make sure that every piece of data made it from AWS to Google Cloud intact,” says Burgher. “We needed to ensure every piece of data stayed complete and linked correctly.”

Salesloft also wanted to avoid downtime. They planned maintenance windows for weekends and late nights to ensure customers didn’t experience any performance gaps during this transition. The sheer volume of data made this a tall order and a very complex task.

The solution: Unifying the environment on Google Cloud

Salesloft knew they couldn’t migrate to Google Cloud alone. They looked for a solutions provider with a history of large moves. They also needed someone who knew Google Cloud technologies inside and out. They chose Insight, which met those qualifications.

Insight brought in expert engineers to join the Salesloft team. They merged into one big team, attended the same daily meetings, and used the same task boards. “The more closely Insight and Salesloft worked together, the better the results were,” says Burgher.

Moving data between clouds smoothly and cost-effectively

Insight used the Google Cloud Database Migration Service (DMS) to help with the move. When they hit a speed bump, such as with PostgreSQL versions, Insight worked with Google Cloud to smooth the process out. The assigned Insight solution architect helped lead the high-level strategy, providing the team with ways to work around blockers.

To move the data, Insight set up a private interconnect between AWS and Google Cloud. The interconnect acted like a VPN on steroids, creating a private pipe for data to travel through while ensuring all data remained within its designated residency boundaries throughout the migration. That was better than using the public internet because it was more secure and far more cost-effective. Insight engineers handled 98% of the work to set up this data pipeline.

The Insight Global Delivery Center kept work going 24 hours a day, matching engineers in India and Armenia with the Salesloft SRE team in the same time zones. Insight built the new system and checked for quality around the clock, which was a key part of making the migration a success.

Strategizing a Big Bang migration

The Insight team also recommended changing the strategy for the final Salesloft move. At first, the plan was to move things slowly without going offline, but this proved too complicated, costly, and time-consuming. In the end, the migration switched to a Big Bang approach, where the platform was unavailable for a short period, and the move happened all at once.

“Originally, we wanted to do the migration with the platform online without any downtime, which made its complexity go through the roof,” says Burgher. “With the increased complexity came increased time and costs. After Insight and my team reevaluated the strategy, a Big Bang migration ended up as the best option.”

The outcome: Achieving scalable growth and speed

Salesloft now lives entirely on Google Cloud, which gives Burgher and team a new sense of peace and clarity. “Take cost tracking as an example,” says Burgher. “We only worry about the hosting costs on a single cloud provider as opposed to two, which translates to the very real convenience of only having to look in one place for our total cloud spend.”

One of the best results was how invisible the move was to Salesloft customers. Even though Insight moved 120 applications and 1 PB+ of data, everything worked nearly perfectly on the first Monday morning after going live. “There were three tickets from customers saying they couldn’t get into the system,” says Burgher. “Salesloft’s customer care team was ready for a busy day, but only one issue came up. It was fixed in hours, but before customers started logging in on Monday.”

Quantifiable benefits

• Moved more than 1 PB of data and 120 apps from AWS to Google Cloud
• Migrated 200 TB of database clusters with full data integrity
  
• Avoided significant costs by bypassing public internet egress fees via a private data interconnect
  
• Slashed time needed to build infrastructure foundations for new features

Qualitative improvements

• Cut cognitive load on SRE and DBE teams by focusing on one cloud
• Accelerated the velocity of the shared responsibility model between SRE and delivery teams
  
• Gave better peace of mind to engineers by simplifying the tech stack
  
• Simplified cost tracking through a single cloud console for better FinOps

Future outlook

Salesloft is now in a great spot for the future. They recently merged with Clari, and because the cloud move is over, the SRE team can spend their time integrating the two companies. Beyond their current toolkit, they’re also looking at new Google Cloud security tools.

Because they only use one cloud, Salesloft can use tools that Google Cloud supports directly. This opens up a larger pool of solutions for the team. They no longer have to find tools that work in both AWS and Google Cloud. This gives them the freedom to innovate without being held back by complex setups.

Secure your path to AI readiness. Explore the four phases of migration and learn how to modernize seven critical workloads with Google Cloud.