SalesLoft Case Study

SalesLoft Case Study - thumbnail.jpg

SalesLoft logo.jpg

SalesLoft’s platform is embedded in its clients’ sales processes, serving as an essential tool for reps and sales leaders who succeed in fast-paced, growth-oriented environments. With VividCortex, SalesLoft was able to decrease its use of temporary tables by 90%, which allowed its engineers to reduce their instance sizes, saving substantial operational costs.

Designed for heavy daily use, SalesLoft’s systems must consistently run at peak performance, so sales professionals can always be ready to close deals and continue to cultivate more. Its environment includes Ruby and JavaScript for development with backend databases that include Postgres, Redis, and ElasticSearch for high volume structured and unstructured search and analytics.

"VividCortex’s Profiler tool for query comparison allows us to immediately rule out infrastructure, hosting provider, and network latency as the root cause of a performance problems."
Mike Shepet, Director, Site Reliability Engineering


Mike Shepet, SalesLoft’s Director of Site Reliability Engineering, is one of the members of the engineering team that keeps SalesLoft’s performance top-notch. He told us that before turning to VividCortex, “The monitoring tool we were using didn’t give us full visibility into our database internals. It couldn’t show us what else was going on in PostgreSQL when our systems started to slow down.” There had been issues isolating poor queries that were hurting system performance.

When the SalesLoft engineering team pushes new features to production, they need to be able to quickly troubleshoot any issues that impact system performance. The performance measurement tools they used in the past were helpful but limited, unable to provide the level of granularity that SalesLoft’s engineering team requires to pinpoint issues. They need precision in order to understand the frequency of various queries, to identify which queries take the most time overall, and to determine where teams should focus efforts to best optimize application performance.

Mike explained that SalesLoft’s engineering team uses a continuous delivery approach, so reducing the mean time to repair database related incidents and changes is critical to minimizing any production issues. “When developers push code to production, they need immediate visibility into the impact of their changes,” Mike said.


SalesLoft began using VividCortex in order to see exactly how code changes affect its database. Mike and his team can look at VividCortex’s Profiler to monitor and compare query performance over different time periods, including before and after code pushes. The engineering team isolates and addresses production issues much faster, frequently identifying and pushing a fix in under 30 minutes. They also take advantage of the AWS CloudWatch integration available within VividCortex, which Mike characterized as, “insanely useful.” 

With VividCortex’s CloudWatch integration, SalesLoft gets a more extensive history of CloudWatch graphs. By going beyond the default two-week retention period, SalesLoft’s engineering team visualizes trends over a more significant period of time. A longer historical view lets the team identify patterns in system utilization, so they can more accurately plan out and optimize system resources. Citing a specific example, Mike said his team had used historical analysis to identify how often they were creating temporary tables that consumed system resources.

Based on this analysis, SalesLoft was able to decrease its use of temporary tables by 90%, which allowed its engineers to reduce their instance sizes, saving substantial operational costs. Just as SalesLoft serves sales teams with the sophisticated tactics they need to close deals, SalesLoft’s data platform requires a sophisticated monitoring solution--VividCortex ensures that Mike and his team can immediately track the impacts of code changes, even at the system’s deepest levels. 


  • Monitor and compare query performance over select time periods.
  • 1-second resolution data.
  • Historical data retention.
  • VividCortex's AWS CloudWatch integration.
  • Intuitive dashboards and visualizations.
  • Substantial savings in operational costs.


"While other products look at different components of the stack, VividCortex fills a key performance monitoring gap by focusing on deep integration with the database." 


Read the full case study here: