Case Study: Intercom achieves customer-level observability and performance insights with Honeycomb

Intercom - Customer Case Study

Intercom, a customer messaging platform for sales, marketing, and support, needed better observability across its monolithic Rails app, Go-based services, and AWS environment. Their existing tools couldn’t handle high-cardinality data like unique customer app IDs, making it hard to pinpoint performance issues or errors affecting individual customers, especially for operations like message matching duration.

With Honeycomb, Intercom added Messenger timing data and used the platform to investigate high-cardinality fields without reworking dashboards or queries. Honeycomb helped them find that one customer accounted for 55% of all server time spent on message matching and more than 10% of overall server CPU usage; this insight came from a 15-second query and also informed scaling and performance analysis for Elasticsearch services.

Intercom

Jamie Osler

Product Engineer