08:00AM – 08:30AM PST
Lounges Open
Engineers are live in the ScyllaDB Lounge and ready to answer your questions!
08:30AM – 08:35AM PST
Welcome & Housekeeping
08:35AM – 9:05AM PST
Session 1: Why Scaling is So Hard
Everything hurts more at scale. We begin by looking at the specific factors that make write-heavy workloads challenging to scale and discussing the related tradeoffs you need to negotiate.
– Write challenges commonly faced in different use cases: IoT, gaming, ad tech, financial
– How database internals and configuration options impact write performance
– Models for predicting your breaking point
9:05AM – 9:35AM PST
Session 2: Strategies for Optimizing Write Performance
There's no single fix for write performance; the right approach depends on your workload characteristics and the bottleneck cause. We cover a range of factors, from data modeling to compression to batching.
– Data modeling and partition design: avoiding hotspots and contention
– Batching: why the relational mental model breaks down at scale
– Other considerations: chunk size, CPU and memory resources, and disk utilization and throughput
9:35AM – 10:00AM PST
Session 3: Troubleshooting Write-Path Bottlenecks
Together, Felipe and Guilherme walk through specific real write degradation scenarios – how to recognize the symptoms, trace them to root causes, and resolve them.
– What to watch for in your monitoring dashboards
– Real examples of write bottlenecks, including symptoms, causes, and remediation strategies
– Configuring custom alerts to warn you of emerging issues
10:00AM – 10:20PM PST
Open Q & A
Guilherme and Felipe answer the audience’s toughest questions regarding real-time write-intensive workloads.
10:20AM – 10:25PM PST