We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Lobsters·Jun 16·5 min read
From Christmas Outage to #1 App Store Ranking: An Aura Frames Postgres Scaling Retrospective
TL;DR
On Christmas Day 2024, Aura Frames experienced a three-hour outage due to Postgres database overload during peak usage. To address this, the team reworked data access and upgraded their Postgres infrastructure, resulting in a system that could handle 226,000 transactions per second.
✦ Why It Matters
Engineers can learn the importance of proactive database scaling and optimization to handle peak loads effectively.
Key Takeaways
- Implement proactive database scaling for peak times.
- Optimize data access patterns to improve performance.
- Monitor transaction rates to anticipate load.
- Test infrastructure under simulated peak conditions.
Full Summary
Aura Frames faced a critical outage on Christmas Day 2024 when their Postgres database could not handle the high traffic, leading to three hours of downtime. To prevent a recurrence, the team restructured their data access patterns and scaled their Postgres infrastructure, which included optimizing queries and increasing database resources.
They achieved a peak of 226,000 transactions per second (TPS) and maintained over 100,000 TPS for ten hours on multiple days. Average query response times improved to just 25 microseconds.
These enhancements allowed customers to set up new frames and upload photos seamlessly, leading to increased app usage. Consequently, the Aura Frames app reached the top position in both U.S. and Canadian App Stores on Christmas Day 2025.
This case illustrates the importance of proactive scaling and optimization in handling seasonal traffic spikes.