Conference Description
Key Takeaways
- Free virtual conference focused on P99 (99th percentile) latency optimization in distributed systems, databases, and cloud infrastructure
- Technical sessions cover kernel-level performance (io_uring, eBPF, QUIC), programming languages (Rust, C++, Go, Zig, Java), and observability at scale
- Case studies from Netflix, AWS, Salesforce, DoorDash, and Snap on real-world performance challenges
- Designed for mid-level to principal engineers, SREs, and system architects working on latency-sensitive applications
About the Event
P99 CONF is a two-day virtual conference taking place October 21 to 22, 2026, organized by ScyllaDB. The event focuses on the P99 metric, which measures 99th percentile latency and represents worst-case performance in production systems. This metric is particularly relevant for engineers building distributed systems, databases, and cloud infrastructure where tail latency directly impacts user experience and system reliability.
The conference takes a vendor-agnostic and tool-agnostic approach, featuring keynotes, technical deep dives, and discussions from practitioners, researchers, and open-source contributors. Sessions run across two half-days, with on-demand video access available after the event.
Technical Topics
Sessions address performance engineering across multiple layers of the stack. Kernel and OS-level topics include io_uring, eBPF, QUIC, and TCP/IP optimization. Programming language coverage spans Rust, C++, Go, Zig, and Java, with attention to how language choices affect latency and throughput.
Infrastructure topics include Kubernetes, serverless architectures, and cloud-native design patterns. Database and storage sessions cover technologies such as Apache Kafka, Apache Flink, Aurora DSQL, Valkey, Redis, TigerBeetle, and JuiceFS. Additional sessions address observability, profiling, debugging at scale, benchmarking, concurrency, and scheduling.
The program also includes content on data formats and processing frameworks including Parquet, Arrow, and DataFusion, as well as infrastructure components like Envoy proxy and the Seastar framework.
Who Should Attend
P99 CONF targets software engineers, backend developers, and system architects working on performance-critical applications. The audience includes performance engineers, SREs, DevOps practitioners, and infrastructure teams responsible for latency-sensitive systems. Database and storage engineers will find relevant content on optimization techniques for high-throughput workloads.
The conference is suited for mid-level to senior and principal engineers at technology companies, cloud providers, SaaS organizations, fintech firms, and teams building large-scale data systems. Technical leads and hands-on CTOs working directly with distributed systems will find applicable content for diagnosing tail latency, scaling databases, debugging complex performance issues, and optimizing resource usage and cost at scale.