Twilio, Github, and Under Armour gain complete visibility with Lightstep

See how!

Customers > Airtable

Airtable Proactively Improves Performance with Lightstep and Reduces Page Load Time by 1 Second for All Users

Airtable’s mission is to democratize software creation by enabling anyone to build the custom work system that meets their needs. Creators and creatives around the world use Airtable to do everything from cattle tracking to filmmaking — and leading brands like Tesla, Airbnb, Slack, Condé Nast Entertainment, and Shopify rely on Airtable to break away from traditional grid-like databases and configure views custom to their use cases.

Airtable’s rapid growth brought massive increases in traffic and significant spikes in peak load.

How could the team both optimize performance over time and ensure rapid restoration of service when issues arise? How could they identify user-facing issues tied to problems deep in their stack, such as database failures?

Through Lightstep, Airtable ensures the reliability of their complex, distributed system — by proactively improving performance for all users, and more effectively monitoring and resolving both systemwide and specific customer issues.

Identifying and Confirming Effective Performance Optimization

Page load times are critically important for Airtable, as their customers expect immediate and easy interaction with their application.

With Lightstep, Airtable was able to identify, measure, and confirm performance optimizations that significantly improved the customer experience.

“Lightstep makes it easy to identify areas of improvement, and then to make sure our optimizations deliver on the intended impact,” said Caleb Meredith, product engineer at Airtable.

Meredith’s team realized there were some things being done serially that could be done in parallel. Since Lightstep provides granular and specific insights into performance, Airtable had visibility into exactly what needed to be optimized.

“We improved "first load" time by 10%, and we were able to quickly perform this work and confirm that it was indeed effective,” said Meredith.

The Airtable team was able to enable the performance optimization for 100% of users in production.

“This is a big win,” said Meredith. “This performance improvement applies to everyone who uses Airtable, no matter how big or small their use case.”

Measuring Anything and Everything

With conventional monitoring, limited cardinality and overly sampled data can make it challenging to gain complete visibility across services.

Lightstep provides Airtable with immediate, real-time insight to the health and availability of any dependency, service, or system component.

“Lightstep shows us exactly where things are slow, what changes we need to make to realize a quick ROI, and specifically how our customers are experiencing our platform,” said Software Engineer Philip Zeyliger.

For example, using Lightstep, the Airtable team was able to measure and reduce the time spent serving linked record searches, bringing p50 down by a whopping 99% — going from roughly 1s to 10ms on average. Additionally, p95 linked record search time was reduced by 62% from 2.1s to 0.8s.

“We’ve made major frontend performance improvements with Lightstep,” said Meredith.

A Better Customer Experience

As a rapidly growing organization, ensuring exceptional system performance is one of Airtable’s primary concerns. Lightstep is Airtable’s first stop when responding to incidents or trying to understand performance within their system.

“Lightstep even enables us to go beyond identifying issues at the specific user level,” said Zeyliger. “Once we’ve conducted an investigation, we can quickly prove whether or not we fixed the underlying issue.”

“Not only does Lightstep help us with understanding the edges of our system, it allows us to monitor and understand new usage patterns that we want to keep a closer eye on to ensure we have an optimal customer experience,” said Doug Forster, Head of Security Infrastructure at Airtable.

“With Lightstep, we’ve been able to scale rapidly and with confidence,” said Zeyliger. “The ability to analyze 100% of our production traffic and leverage high-cardinality insights has dramatically improved how we monitor, optimize, and automate our system.”

“Lightstep has helped us make major strides in ensuring an exceptional experience for every Airtable customer.”

Developer shortcut

Learn how you can share your findings through Slack with Lightstep snapshots

View the docs
Airtable
Headquarters:

San Francisco, CA

Industry Segment:

SaaS

Employees:

200+

Funding:

$170M

Challenges

  • Limited visibility into specific end user performance
  • Incomplete understanding of system performance
  • Uncertainty around the impact of performance optimizations

Results

  • Customer-specific high cardinality monitoring
  • Complete real-time system visibility
  • Ability to identify and confirm the effective performance optimizations
"Lightstep makes it easy to identify areas of improvement, and then to make sure our optimizations deliver on the intended impact."

Caleb Meredith
Product Engineer at Airtable
"Not only does Lightstep help us with understanding the edges of our system, it allows us to monitor and understand new usage patterns that we want to keep a closer eye on to ensure we have an optimal customer experience."

Doug Forster
Head of Security Infrastructure at Airtable
"Lightstep shows us exactly where things are slow, what changes we need to make to realize a quick ROI, and specifically how our customers are experiencing our platform."

Philip Zeyliger
Software Engineer at Airtable

Key takeaways

Challenges

  • Limited visibility into specific end user performance
  • Incomplete understanding of system performance
  • Uncertainty around the impact of performance optimizations

Results

  • Customer-specific high cardinality monitoring
  • Complete real-time system visibility
  • Ability to identify and confirm the effective performance optimizations

Next case study:


Digital Ocean

Digital Ocean

DigitalOcean uses Lightstep as a Source of Truth for its Distributed System, Saving 1000 Hours of Developer Time per Month