P99CONF

Thanks to new features of the #Linux kernel, anyone can write a basic Linux scheduler and use it to fuzz concurrency bugs or optimize for specific workloads. Jake Hillion and @parttimen3rd showed how to do it at #P99CONF. Watch it free here > ow.ly/E9He50XmveV #ScyllaDB

Durable workflow engines typically depend on adding overhead, write amplification, and complexity. At #P99CONF, Qian Li shared a lightweight library-based engine that embeds into application code and checkpoints state directly to the database. ow.ly/uBSP50XmvKr #ScyllaDB

pgBadger's "raw output" feature allows you to build new tools to root out the causes of performance degradation. In a highly technical #P99CONF session, @HettieDombr explained how to uplevel pgBadger so you can unlock deeper insights. ow.ly/wNM950XJC4P

The cost and efficiency of IP geolocation data services depend on careful system design. Yichen Wei explained how to enable higher traffic with lower resource use in a highly technical #P99CONF session. Watch it for free here: ow.ly/KHnv50Xq2TL #ScyllaDB

Tokio is always a hot topic at #P99CONF, and we're looking forward to seeing our friends @glcst, @carllerche, and @problame at @TokioConf 2026. You can get more details here. tokioconf.com

If you're moving from C++ to #Rustlang, @ClickHouseDB, an open source analytics-oriented database system composed of 1.5 million lines of code, has indispensable advice. Check out what Alexey Milovidov shared at #P99CONF via @thenewstack. ow.ly/ce0q50XqAHX #ScyllaDB

Maximizing the KV Cache hit rate is critical to minimize time-to-first-token latency. In John Thomson's #P99CONF talk, the techniques NVIDIA’s TensorRT-LLM uses to maximize KV Cache hit rates for structured LLM workloads was shared. Watch here > ow.ly/SIue50Xm5Ps #ScyllaDB

After two days of technical talks at #P99CONF, we're enjoying reading the takeaways from our attendees. Dhruv Pandita was kind enough to look at the four things forcing a radical, full-stack reckoning. ow.ly/fzrM50Xiq6n

How was elfo transformed into a distributed mesh of services? At #P999CONF, Nikita Lapkov explained the transformation and how they made everything observable and debuggable. Watch it for free here > ow.ly/X02g50Y0LOm #ScyllaDB #Rustlang

Spoiler alert - our intuition on where to find performance issues is usually wrong. At #P99CONF, AWS' Geoffrey Blake shared APerf and an example of how it debugged a difficult P99 latency problem. Watch it for free here > ow.ly/pF9M50Y7KRx #ScyllaDB

What's the best way to compare @ScyllaDB and #Memcached? Fair benchmarking can be difficult to come by, so they teamed up to do it together! Check out the results in this #P99CONF on-demand session. ow.ly/TEXm50UgUlb #ScyllaDB #lowlatency #database

One of #P99CONF's most highly rated speakers is @tanelpoder. Tanel shared the prod/GA ready version of xCapture v3 + xtop and demonstrated how it helps with real life troubleshooting scenarios. Watch the entire session here for free: ow.ly/QfGR50XpsSN #ScyllaDB

Async programming can be tricky, but the right tinkering and visualization can make it click. At #P99CONF, Alex Puschinsky's built an async #Rustlang visualization tool to see what happens when CPU-heavy code sneaks in. Watch for free here > ow.ly/kPKE50Ycli9 #ScyllaDB

Spoiler alert - our intuition on where to find performance issues is usually wrong. At #P99CONF, AWS' Geoffrey Blake shared APerf and an example of how it debugged a difficult P99 latency problem. Watch it for free here > ow.ly/pF9M50Y7KRx #ScyllaDB

Kenny Chamberlin's #P99CONF session covered reducing thread counts to cut tail latency, scoping locks tightly, 7 removing locks with help from Rust’s borrow checker. Watch for free to learn practical examples with metrics, code, & live demos. ow.ly/lEne50XpsFn #ScyllaDB

One of #P99CONF's most anticipated talks was from @business' Kusha Maharshi. Kusha showed how they wrestled complicated DAGs into shape, thus generating end-to-end latency metrics that drive SLOs, fire off alerts, & keep them honest. Watch here: ow.ly/Vyv350XjacT #ScyllaDB

Saurabh Jaluka and Tulika Bhatt introduced Netflix's API-driven system that centralizes content and module selection, decouples business logic from clients, and standardizes data models for faster innovation at #P99CONF. Watch it here: ow.ly/thzU50XlYwA #ScyllaDB

Observability lets your team know the system's status without checking under the hood. From QA processes to daily life, @SrPerf's #P99CONF session explains how its principles can enhance awareness and efficiency. Check it out here > ow.ly/Az6W50UbaYq #ScyllaDB

The use of #Rustlang in the kernel allowed for the implementation of a lot of safeguards. #P99CONF alum Luc Lenôtre explains why switching from C language to Rust was the best possible decision for a project. ow.ly/k15q50TfjHW #ScyllaDB

LLM deployments require massive GPU demands and costs. One way to solve this is by using a generic architecture for offloading KV-cache tensors to a disaggregated shared store. Watch Eshcar Hillel's @P99CONF session to learn more > ow.ly/87ww50XR35r #ScyllaDB #P99CONF