Stenobird

Episode

From Aurora to PlanetScale: Intercom’s Database Evolution with Brian Scanlan

Podcast
Screaming in the Cloud
Published
Sep 18, 2025
Duration seconds
2613
Processing state
processed
Canonical source
https://share.transistor.fm/s/cc385786
Audio
https://dts.podtrac.com/redirect.mp3/media.transistor.fm/cc385786/2f78ee10.mp3
JSON
/v1/public/podcasts/screaming-in-the-cloud/episodes/from-aurora-to-planetscale-intercom-s-database-evolution-with-brian-scanlan
Markdown
/podcast/screaming-in-the-cloud/from-aurora-to-planetscale-intercom-s-database-evolution-with-brian-scanlan.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/screaming-in-the-cloud/episodes/from-aurora-to-planetscale-intercom-s-database-evolution-with-brian-scanlan/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/screaming-in-the-cloud/from-aurora-to-planetscale-intercom-s-database-evolution-with-brian-scanlan.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Intercom transitioned from AWS Aurora to PlanetScale to solve the operational complexity of managing 13 separate database clusters. The discussion explores why specialized providers are outperforming AWS by offering superior developer experience and white-glove support.

Topics

  • AWS Aurora
  • PlanetScale
  • Vitess
  • Database Sharding
  • Cloud Infrastructure
  • AI Chatbots
  • Developer Experience
  • Site Reliability Engineering
  • Snowflake

Highlights

  • Main idea: Specialized providers like PlanetScale and Snowflake are winning by offering superior user experiences and faster issue resolution compared to AWS
  • Failure mode: Relying on AWS's 'building-block' approach can lead to massive operational overhead, as seen with Intercom's 13 Aurora clusters
  • Practical takeaway: High-quality managed services (e.g., via Slack integration) allow engineering teams to resolve critical incidents much faster than traditional ticketing systems
  • Main idea: The shift toward AI agents at Intercom is driven by the need to move beyond 'bad implementation' of chatbots to truly useful automated support
  • Practical takeaway: Engineering culture should prioritize human-centric on-call rotations to prevent burnout and maintain long-term talent retention

Chapters

  1. 4:10 The AI Chatbot Revolution: Intercom's pivot toward advanced AI agents following the explosion of LLM technology.
  2. 10:40 Scaling Challenges with Aurora: The technical necessity of moving away from Aurora and building custom sharding solutions.
  3. 17:20 Why PlanetScale Won: Comparing the managed Vitess experience of PlanetScale against the operational pain of AWS clusters.
  4. 20:30 The Rise of Specialized Providers: How Snowflake and PlanetScale are outperforming AWS by focusing on the application layer and customer support.
  5. 23:45 The Cost of Enterprise AWS Support: The friction caused by expensive, unbounded enterprise support models and the difficulty of navigating AWS's support ecosystem.
  6. 37:00 Human-Centric Engineering Culture: Designing on-call rotations and engineering processes that prioritize developer well-being and human needs.