Saint HighLoad++ 2025
2025-06-23..2025-06-24
-- online --
[Saint HighLoad++ 2022, Saint HighLoad++ 2023, Saint HighLoad++ 2025]
| title | author | length |
|---|---|---|
| Opening of Saint HighLoad++ 2025 | Dasha Bykova | 15m |
An opening presentation for the Saint HighLoad++ 2025 conference presenting the professional event for developers of high-load systems. It gives logistic… | ||
| How to Choose Technologies for a High-Load Project | Dmitriy Krivopaltsev, Vadim Kleba | 48m |
Guidance on choosing programming languages, databases and other technologies for high‑load services, and how to design a long‑lived high‑load application… | ||
| When You Can't Avoid a Seq Scan: Data Skipping in the New Tarantool Columnar Engine | Andrey Saranchin | 48m |
The talk explains data skipping as an alternative to full table (Seq) scans when precise indexes aren’t available, focusing on its implementation in… | ||
| Seven Circles of Fintech: A Drama in Two Acts | Margarita Monogarova | 37m |
An exploration of practical challenges in building fintech payment integrations, contrasting the idealized view (good docs, reliable providers) with real-world… | ||
| The Evolution of Vector Search in YDB | Aleksandr Zevaykin | 33m |
Describes YDB’s development of vector search from a simple initial implementation to an advanced indexing system, detailing the key development stages… | ||
| Lego Stand: The Story of Building a Universal Test Lab | Evgeniy Derkach | 42m |
Describes building a unified remote test lab from 30+ heterogeneous devices (IP cameras, smart meters, tablets, single-board computers, etc.) with varying… | ||
| How to Save Gigabytes of Memory in Istio Sidecars | Maksim Chudnovskiy | 51m |
Examines memory usage of Istio sidecars in large and hyperscale production deployments, showing that a single sidecar can reach gigabytes of RAM rather than… | ||
| ML in Production: Why Analysts and Backend Struggle to Agree | Sergey Lavrov | 40m |
The talk describes how Aviasales deployed ML scoring in production to rank flight tickets by purchase probability. It covers practical issues encountered—data… | ||
| The Evolution of Database Architectures | Vladimir Komarov | 51m |
Survey of how traditional enterprise DBMS architectures (Oracle, MS SQL Server, PostgreSQL) — whose core designs date back over thirty years — show… | ||
| CDC Without Pain: real-world resilient replication with Debezium and Kafka | Evg. Prochan | 38m |
An explanation of using Change Data Capture (CDC) with Debezium and Kafka to build resilient replication between systems. Describes practical problems of data… | ||
| MLOps Architecture for a Recommendation Platform | Dennis Khassan | 37m |
A presentation describes an MLOps architecture used to deploy a recommendation platform across Sber’s ecosystem, covering services from streaming to… | ||
| Uniting Complex Branch Organizations with an Event Mesh | Aleksandr Bardash | 49m |
Describes integrating complex subsidiary organizations using an Event Mesh to connect heterogeneous tech stacks and maturity levels. Explains how to bridge… | ||
| High‑Load RAG Architecture: 10 Chunking Optimization Strategies | Andrey Nosov | 53m |
Architectural solutions for high-load retrieval-augmented generation (RAG) systems where document chunking is a scaling bottleneck are examined. Ten concrete… | ||
| How to Build Video Analytics for Live Streams Without Big Tech Resources | Daniil Blinov | 38m |
A team from Sibur describes building video analytics for live streams from over 2,000 cameras across multiple factories, tackling tasks like production process… | ||
| Architectural Quiz: Hack or Elegant Solution? | Natalya Makarova | 51m |
Examines practical architectural decision-making for building resilient, scalable enterprise systems, arguing that real-world context matters more than… | ||
| Distributed Mapping Technologies for Autonomous Vehicles | Gosha Lutidze | 39m |
Describes building a distributed map system for autonomous vehicles that delivers high accuracy, reliability, and scalability. Covers the technologies and… | ||
| Production Traffic Collection Architecture for Load Testing | Evgeniy Kuzyshin | 49m |
Describes how Ozon collects production traffic for load testing by cloning real user requests from other services; the architecture handles aggregating and… | ||
| Spark: One Tool for Everything? | Iskander Fakhrutdinov | 46m |
The talk examines Apache Spark's role as a general-purpose data processing engine after more than a decade of changes in hardware, storage formats, and… | ||
| How We Built a Unified Ecosystem Customer Profile at MTS | Evgeniy Nenakhov | 52m |
MTS built a unified ecosystem customer profile that aggregates data across multiple industries (telecom, fintech, carsharing, streaming, music, travel, etc.)… | ||
| Schrödinger’s Warehouse: Reducing Incidents with Computer Vision | Mikhail Krasilnikov | 50m |
The talk describes using computer vision to reduce incidents in large logistics centers by detecting human errors, procedural violations, and theft from… | ||
| Infrastructure Provider for Cluster API: From Zero to Open Source | Vasiliy Dimov | 34m |
Describes building an infrastructure provider for Cluster API from scratch and releasing it as open source. Covers using Cluster API to declaratively manage… | ||
| Automatic Summarization of 10K Daily Meetings: from Requirements to Production | Azer Shakhverdiev | 49m |
A team that runs a video-conferencing product creating over 10,000 meeting recordings per day explains how they moved from a one-hour SLA for automatic meeting… | ||
| Scaling Limits of the Sokol Disk DBMS | Andrey Korotchenko | 49m |
Examines scaling limits of a disk-based DBMS called Sokol, evaluating whether a pure disk-only database (without caching front layers) can meet practical… | ||
| 20 Years on the Rake: Failures, Outages and Lessons Learned | Aleksandr Sterligov | 53m |
A 20-year retrospective on production mistakes and outages in software development and operations, covering memorable failures, surprises, and lessons learned… | ||
| How to Cook RabbitMQ Right — 8 Practical Cases | Viktor Mikhaylov | 52m |
Practical guidance on using RabbitMQ in production, focusing on eight real-world cases that show how to avoid losing messages. Covers RabbitMQ-specific… | ||
| "Manual" Vectorization in 2025: When the Compiler Is Not Enough | Valeriya Puzikova | 51m |
Examines manual SIMD/vectorization techniques on RISC-V and Arm using Eigen and OpenBLAS as case studies, comparing their performance to what compilers'… | ||
| Thousands of Async Tasks per Second in the Cloud on Rust/Axum/Tokio | Aleksandr Serbul | 51m |
Describes building Rust command-line utilities using Axum and Tokio to perform massive asynchronous, non-blocking operations against AWS-compatible S3 clouds… | ||
| How We Built Our Hardware Solution for Measuring Product Dimensions and Weight | Mariya Gafurova | 49m |
A team built a custom hardware system to measure product dimensions and weight for automated logistics. The talk describes why businesses need fast, accurate… | ||
| 10K Metrics, 500 A/B Experiments and 500M p‑values Every Day — How Is That Possible? | Danyla Lenkov | 47m |
An in-house A/B experimentation platform built at Avito is described, explaining how it scaled from a 2018 prototype to a production system running about 4,000… | ||
| Migrating an Ad Engine to the Cloud from Bare Metal Under High Load | Artem Bukin | 47m |
A team migrated a high-load banner ad engine from bare-metal servers to a cloud environment. The talk describes goals for the move, the problems encountered… | ||
| JOINs Are Slowing You Down: Why Spark and Trino Won’t Replace Your DWH | Aleksey Dmitriev | 47m |
Examines why JOIN-heavy workloads can be slow and why query engines like Apache Spark and Trino may not be suitable substitutes for a dedicated data warehouse… | ||
| SLI and SLO for the Business: Monitoring Quality Across 200+ Products | Filipp Bocharov, Eduard Stepanyants | 57m |
MTS describes how they moved from host- and app-level monitoring to tracking SLI/SLO for full business scenarios across 200+ (400+ mentioned) products. The… | ||
| How PG Sharding Works in Yandex Taxi’s Processing | Igor Bereznyak | 50m |
Describes Yandex.Taxi’s custom PostgreSQL sharding solution inside its order processing core, including how data is organized to meet high load, strict… | ||
| Porting ML to RISC‑V: How Not to Lose Performance | Kseniya Zaytseva | 36m |
Examines porting machine-learning workloads to RISC-V processors, focusing on performance issues caused by limited bus bandwidth when using tensor or GPU… | ||
| A Million Products, One Experience: Using Collaborative and Multimodal Embeddings | Oleg Durandin | 52m |
Describes building a product clustering system for a marketplace to create thematic collections and improve user experience for millions of items. Explains… | ||
| Exploratory Testing: How to Find Backups That Won't Restore | Aleksandr Fatin | 46m |
The talk explains exploratory manual testing in large high-load products and argues it is a sign of engineering mastery. It shows how a mature approach to… | ||
| How We Built a Unified Online Tariffing Platform | Nazar Geldyev | 36m |
Describes building a unified online billing/tariff platform for an operator-grade, high-load system serving over 80 million subscribers. Covers how the team… | ||
| Analytics on Large Graphs in S3: Tools, Approaches and Formats for OLTP and OLAP | Aleksey Teplov | 48m |
Examines tools and formats for high-performance analytics on very large graphs stored in S3, focusing on workflows for OLTP and OLAP use cases. Describes… | ||
| ClickHouse Without Pain: Our Experience Refactoring a Legacy Cluster and the Impact of Shard Groups | M.Filimonov | 47m |
Practical lessons from refactoring a legacy ClickHouse cluster: common mistakes in building and operating ClickHouse, the negative consequences observed, and… | ||
| Are Four Nines Enough for Everyone? Understanding Data Center Availability | Aleksey Uchakin | 45m |
The talk examines data center reliability claims by inspecting uptime metrics, TIER classifications, and SLA 'nines' (e.g., 99%, 99.9%, 99.99%, 99.999%). It… | ||
| How We Balance Load Inside Yandex Messenger | Nikita Zvonarev | 47m |
Describes how Yandex Messenger balanced load across nodes of a key stateful service by eliminating manual rebalances and reducing node failure background rates… | ||
| Who Wrote the Code? On Copyright for Code Written with AI | Irina Shakhtarina | 48m |
Explains who typically owns copyright in software and examines how that changes when code is generated with AI assistants. Covers legal issues hidden in user… | ||
| Dynamic Ranking of Search and Recommendation Results in High-Traffic Systems | M.Panenko | 41m |
Describes a system for dynamic ranking of search and recommendation results in high-load job platforms to increase employers’ vacancy views by 3–10× without… | ||
| The Technical Evolution of Auto.ru Classifieds: From the Beginning to Today | Andrey Borunov | 50m |
A chronological overview of the technical evolution of the Auto.ru classifieds site from its 1996 origins as a simple bulletin board and forum to its current… | ||
| Real-Time Supply Chain Analytics Pipeline at X5 Based on Debezium CDC and Flink | Ivan Smolyaninov | 34m |
Describes a real-time supply-chain analytics pipeline built at X5 Group using Debezium for change-data-capture and Apache Flink for stream processing. Covers… | ||
| Bypassing LLM Defenses with Adversarial Suffixes and AutoDAN | Nikita Belyaevskiy | 53m |
The talk explains what adversarial (competing) suffixes are and demonstrates that any large language model can be vulnerable to them. It shows how to construct… | ||
| Geo-Distributed Postgres Replication Using Debezium | Nikolay Golubev | 51m |
Shows how to use Debezium and Kafka to replicate PostgreSQL data between geographically distributed database nodes, and how to automate failover and minimize… | ||
| MinIO: Scaling and Operations | Aleksey Pletnev | 50m |
A presentation about operating and scaling an S3-compatible distributed object store built on MinIO, sharing ongoing real-world deployment and management… | ||
| Deep Dive into Kafka Architecture: From Simple Scenarios to a Geo-Cluster | Andrey Komyagin | 57m |
A technical deep dive into Apache Kafka’s architecture covering basics through advanced deployment patterns, including single-cluster and geo-cluster setups… | ||
| The DDD Aspect and GraphQL in Building High-Load Integration Systems | Viktor Biryukov | 52m |
Describes using Domain-Driven Design (DDD) together with GraphQL to build high-load integration systems, focusing on reducing backend development effort while… | ||
| SPDK Under the Hood: Diving into the Disk Subsystem in User Space | Vasiliy Ivanov | 48m |
Explains what SPDK (Storage Performance Development Kit) is and why to use it for delivering guest I/O from virtual machines to storage systems. Describes an… | ||
| How Car Engineers Set Up a Technical Race Among Themselves | Ilya Starikov | 52m |
The talk examines how modern cars have become complex, high-load engineering systems and describes a technical race among automotive engineers that prioritized… | ||
| A Custom File Store for 350 PB of Video Content | Edgar Voskresenskiy | 44m |
Description of a proprietary file storage system used to host over 250–350 PB of video for RUTUBE (400+ million items). Explains the custom storage… | ||
| Workshop: Amazing Embeddings — How Vector Representations Are Applied | Ivan Krasnikov | 01h 43m |
Workshop presenting how vector embeddings are used across applications like search engines, recommender systems, business analytics and marketing; covers… | ||
| Masterclass: Let's Cut the Monolith | Aleksey Losev | 01h 52m |
Hands-on masterclass on breaking a large monolith into smaller services, covering when to split (legacy growth or a service becoming too large), and practical… | ||
| A Look Behind the Scenes of Yandex.Maps: From Raw Satellite Data to Clear Images | Igor Orpanen | 46m |
A behind-the-scenes look at how raw satellite imagery is processed for Yandex.Maps, covering the pipeline from initial satellite data through geospatial… | ||
| Comparing the Scalability of Kafka and YDB Topics | Andrey Serebryanskiy | 51m |
Compares scalability of Apache Kafka versus YDB Topics for high-throughput messaging systems, focusing on how each broker behaves when scaling. Describes… | ||
| Modern Mechanisms for Determining Device Parameters on the Web | Evgeniy Sharamed | 42m |
Describes modern techniques for detecting device parameters in web systems amid reduced browser-supplied headers and evolving device models. Covers challenges… | ||
| OVN: Integrating an Off-the-Shelf SDN into a Virtualization Platform | Dmitriy Skorykh | 51m |
Describes integrating OVN (Open Virtual Network) as a ready-made SDN solution into a server virtualization platform to provide virtualized network… | ||
| How companies can share data without sharing it | Petr Emelyanov | 48m |
Explains approaches for protecting data while it’s being used so companies can collaborate without sharing raw data. Covers information security vs. privacy… | ||
| How we built observability on Open Source at ecom.tech: OpenTelemetry, Qryn | Valeriy Evdokimov | 36m |
A practical look at building observability for an open-source e-commerce platform using OpenTelemetry, Qryn and Coroot. Covers telemetry formats… | ||
| Global web resource load balancing across geo‑distributed infrastructure | Dmitriy Krikov | 52m |
Explains how to design global load balancing for large web services using geographically distributed infrastructure to keep sites fast and available worldwide… | ||
| Data sources for fuzzing web application APIs | Vsevolod Dergunov | 49m |
Fuzzing web-application APIs to find vulnerabilities missed by SAST or manual review, demonstrating a concrete example. Focuses on the main challenge of… | ||
| Ranking factors architecture in Ozon search runtime | Lev Saskov | 50m |
Describes Ozon’s runtime ranking architecture that uses models and feature factors for search, including how ranking models are served (inference) in the… | ||
| Optimizing TCO of K8s clusters on AWS and YC | Andrey Ivakhnenko | 39m |
Describes how Antiplagiat reduced Kubernetes TCO on AWS and Yandex Cloud by using autoscaling to handle highly variable daily and seasonal traffic (up to 10x… | ||
| Making low‑code execution transparent: lessons from a large IoT platform | Vladimir Plizga | 47m |
A large IoT platform’s engineers examine making low-code execution transparent by exposing what low-code/no-code constructs do under the hood and how they… | ||
| The endless war in memory: security without slowdowns | Sergey Ignatov | 45m |
Examines hardware-based techniques for preventing memory corruption in high-load systems, following an earlier focus on software defenses. Covers concrete… | ||
| What it takes for stream data processing | Maksim Buylin | 53m |
The talk explains what a full streaming data platform needs: durable event storage, reliable delivery of those events to downstream systems, and stateful… | ||
| Database sharding at the load balancer level | Nikolay Prokofev | 52m |
Describes how T-Bank shard a high-load customer-data service with tables of billions of rows by implementing sharding at the load-balancer layer in Kubernetes… | ||
| On hardware everyone is happy differently, in clouds everyone suffers the same. What to do? | Fedor Petryaykin | 34m |
Compares running high-load applications on bare metal versus in cloud environments, identifying common causes of higher compute resource usage in clouds… | ||
| How business requirements shape architecture: evolution of the notification service at Lamoda Tech | Aleksey Sitka | 51m |
The talk describes how Lamoda Tech’s notification service architecture evolved in response to concrete business requirements, showing which requirements (like… | ||
| How we generate 130,000 shorts per month using AI | Aleksandr Sokolov | 37m |
Describes an automated pipeline that uses AI to generate 130,000 short vertical videos per month by detecting and extracting engaging moments from source… | ||
| Georgy Melikov, Alexander Popov VK, VK Cloud | Georgiy Melikov, Aleksandr Popov | 01h 58m |
The talk examines how container networking is actually built on long-established standard technologies and why people forget that fact. It focuses on the… | ||
| A streaming correlator handling over 2M EPS | Aleksey Kiselev, Andrey Borisov | 41m |
Describes a streaming correlator built to process over 2 million events per second (EPS) for modern Security Operation Centers handling hundreds of systems and… | ||
| Web scraper developer (53 bots) 500m from you and wants to meet | Arseniy Savin | 53m |
An overview of web scraping techniques and countermeasures: describes what web scraping is, common tools, how to detect when a site is being scraped, and both… | ||
| From 1k to 180k IoT devices: how we scaled the architecture | Ruslan Yagafarov | 48m |
A MTS Urent team member describes scaling their backend as the scooter fleet grew from 1,000 to 180,000 IoT devices. The talk covers handling increasing data… | ||
| How to tune Nginx to survive a DDoS | Sergey Cherkashin | 58m |
Practical guidance on configuring Nginx to detect and survive DDoS attacks, covering systemd/systemctl tweaks, monitoring options, and module choices. It… | ||
| Scaling the impossible: how we built an OTA update system | Petr Rastegaev | 51m |
Describes building a custom OTA firmware-update system for thousands of devices across distributed warehouses when commercial cloud solutions were unusable… | ||
| AsyncAPI – Swagger for asynchronous APIs | Aleksandr Mitin | 44m |
AsyncAPI is presented as a specification analogous to Swagger but for asynchronous, event-driven systems. It covers using AsyncAPI to describe interactions… | ||
| Training GigaChat MAX | Evgeniy Kosarev | 43m |
Presentation describes GigaChat MAX, a large Russian-language LLM launched in 2025; it combines state-of-the-art distributed training technologies and curated… | ||
| Dark security patterns: when protection gets in the user's way | Artem Bachevskiy, Natalya Protsenko | 49m |
Examines how security UX choices create 'dark patterns' that make protection confusing or annoying for users. Discusses examples of security features and… | ||
| Automated algorithmic trading bot | Viktor Delisov, Ilshat Khafizov | 50m |
Describes building an automated algorithmic trading bot that manages thousands of advertising campaigns on Telegram using a data-driven approach to maximize… | ||
| Solving real business cases with ClickHouse | Dmitriy Cherobaev | 49m |
Using ClickHouse to solve real business cases: describes problems encountered when building a service with ClickHouse—difficulty with updates, huge data… | ||
| Smart search over internal knowledge base with LLMs: from architecture to deployment | A. Boltava | 56m |
Describes building an intelligent internal knowledge-base search using large language models (LLMs) and retrieval-augmented generation (RAG) to speed up work… | ||
| How We Built a 30+ PB Storage for AI and Plan to Keep Growing | Aleksandr Nepochatykh | 43m |
Describes building a data storage system holding over 30 petabytes for AI workloads, covering requirements, architecture choices, and operational practices… | ||
| Research Under the Microscope: How the Quality Lab Helps Improve VK Video | Aleksey Shpagin | 48m |
Describes VK Video’s Quality Laboratory and how it improves playback for millions of users by diagnosing and optimizing the video platform. Covers the lab’s… | ||
| Passwordless in PAM: Myth or Reality? | Artem Nazaretyan | 47m |
Discussion of integrating passwordless authentication into PAM-based systems, covering technologies such as FIDO2/WebAuthn (hardware keys), TOTP/push one-time… | ||
| How We Sharded and Migrated a 20 TB MongoDB Without Downtime | Ivan Perevoshchikov | 37m |
Avito needed to move multiple sharded MongoDB clusters (each 1–5 TB, totaling ~20 TB) from large fast bare-metal servers to smaller slower Kubernetes nodes… | ||
| From Overheating to Success: How We Deployed Liquid Cooling on Desktop CPUs | Grigoriy Pradedov | 44m |
A team describes deploying custom closed-loop liquid cooling (CLC) for racks of Intel Core i9-9900K desktop CPUs after air cooling failed in 1U server cases… | ||
| Building a Unified Authentication Platform Using Keycloak | Viktor Popov | 46m |
Describes building a centralized authentication platform for a group of companies using Keycloak. Explains the decision to separate authentication and… | ||
| Valkey: Latency in Cloud Environments | Evgeniy Dyukov | 44m |
Valkey (an open-source Redis fork) is examined in cloud deployments where adding a caching layer can actually increase latency because of network hops between… | ||
| From Click to Publish: Architecture of Avito’s Classifieds Platform | Evgeniy Tolmachev | 53m |
Overview of Avito’s ad-publishing pipeline: how a listing goes from the user clicking “post” to appearing in search under heavy load. Covers the platform… | ||
| A 20‑Minute Adventure: Issues and Patches During OK Streaming Upgrade | Alina Shestakova | 32m |
Describes how Odnoklassniki's (OK) streaming system is architected and recounts the concrete problems, patches, and failures encountered during a framework… | ||
| Extension API Server in Kubernetes: When CRD Is Not Enough | Yaroslav Borbat | 40m |
The talk explains when Kubernetes Custom Resource Definitions (CRDs) are not enough and presents using an Extension API Server as an alternative. It covers… | ||
| Turning Off‑the‑Shelf Tools into a Petabyte‑Scale System | Pavel Lukyanov | 45m |
A practitioner describes building a petabyte-scale data system using existing (import-substituted) tools for collection, processing, and storage, including how… | ||
| A Universal Document Index: How We Index Billions of Documents | Evgeniy Redko | 45m |
Explains how Elasticsearch works and scales and why its built-in features are insufficient for all search and filtering scenarios. Describes the motivation for… | ||
| Evolution of IVI’s Home Page, Video Verticalization and Personalization | Dmitriy Gonchar | 40m |
Describes how a major Russian streaming service redesigned its homepage from a static page to a dynamic, personalized video feed. Covers the vertical video… | ||
| How We Rewrote Authentication in a Microservices Platform | Dmitriy Larionov | 44m |
Describes replacing an old corporate SDK and SSO agent in MTS’s partner-relations platform with OAuth2/OIDC. Covers refactoring the platform’s internal… | ||
| Applied Consensus: Which Station Should Respond? | Pavel Korozevtsev | 52m |
Describes a backend streaming algorithm that selects exactly one smart device (a "Station") to respond when multiple devices detect a voice assistant wake… | ||
| A Homegrown S3 Server: Challenges of Building Stable Storage | Maksim Kharitonov | 50m |
Describes Ozon's migration of object storage to their in-house S3-compatible server Lusca, covering choices of index storage technologies, techniques used to… | ||
| Evolution of the PostgreSQL Storage for Listings at Avito | Evgeniy Konstantinov | 51m |
A postmortem-style case study of Avito’s high-load PostgreSQL storage for listings, describing how the team handled large growth in traffic and data without… | ||
| Maintenance #HowInSber: Java DBMS in Critical Systems | Ilya Shishkov | 46m |
Describes development and operational experience with a Java-based DBMS called Platform V DataGrid (Apache Ignite SberEdition) used in over 200 systems and… | ||
| Evolution of Data Collection in the Datalake at Samokat Infrastructure | Dmitriy Koprov | 37m |
Describes implementing EventStreaming to feed a DataLake at ecom.tech: why they adopted event streaming, which problems they aimed to solve, and why they… | ||
| Ethereum 2.0 Network Consensus: A Million Master Replicas, No Jokes | Sergey Prilutskiy | 51m |
Examines Ethereum 2.0’s network consensus as a massively decentralized database, focusing on how over a million active validators operate and periodically act… | ||
| Supply Chain from SLSA to OSC&R | Lev Khakimov | 51m |
Overview of supply-chain security for software delivery, comparing the SLSA framework with other measures and gaps it leaves against certain attack types… | ||
| So Let’s Build a Data Indexing System | Aleksandr Parkhomenko | 42m |
Describes 2GIS’s indexing system for search data that needs very frequent updates (some segments every 10 minutes). Explains how they deliver fresh data… | ||
| AdGuard DNS — From Zero to 100,000,000 Users Through Pain | Andrey Meshkov, Konstantin Zamyakin | 50m |
AdGuard DNS’s eight-year engineering story: building a global DNS service that now handles 5 million RPS, 100 million active users and 10 PB/month. Covers… | ||
| On WAL Redundancy in Postgres | Andrey Borodin | 52m |
Explains redundancy in PostgreSQL WAL (write-ahead log) files and shows that WAL data compresses highly, indicating excess redundancy. Demonstrates how… | ||
| The Cycle of Upgrades | Aleksandr Churikov | 46m |
A talk about updating enterprise storage systems that must remain highly available and continuously operational. It discusses challenges of applying periodic… | ||
| Managing Resources as a Product | Mariya Vasileva | 46m |
Yandex Verticals faced a hard problem controlling resource consumption across thousands of services and hundreds of data stores, and built an internal… | ||
| How to Extract SLOs: Sources and Tools of Midband SRE Gnomes | Dmitriy Sinyavskiy | 44m |
Explains practical ways to obtain and implement SLOs (Service Level Objectives) for teams that already understand SLI/SLO concepts. Covers sources of SLO… | ||
| Code Analysis: A Symbolic Virtual Machine | Georgiy Aleksandriya | 51m |
The talk explains symbolic execution implemented via a symbolic virtual machine for code analysis: it describes treating inputs as symbols to explore all… | ||
| Application and Supply-Chain Attacks: How to Keep Using APIs and Sleep Well | Lev Paley | 49m |
An overview of API and supply-chain attacks with statistics on attack vectors, comparing positive (allow-list) and negative (block-list) protection models and… | ||
| From Production Downtime Due to a Network Error to a Commit in Upstream Linkerd | Vadim Makerov | 31m |
Talk describes diagnosing a production downtime caused by a network bug in a service mesh and the process of fixing it by contributing a patch upstream to… | ||
| How to Build Effective Dashboards for 2,000+ Microservices? | Aleksey Zolotukhin | 49m |
The talk describes designing effective dashboards for over 2,000 microservices in a city services platform that runs Yandex.Taxi, Food, Lavka and Delivery. It… | ||
| A Real-Time Analytics System for 5 Billion Events a Day Using ClickHouse | Andrey Berezin | 50m |
Describes SaluteEye, a product analytics platform that ingests events from Sber smart devices and other sources and provides analytics and monitoring of… | ||
| Infrastructure Satisfaction Metrics: Forgive and Understand or Let Go? | Aleksandr Krylov | 52m |
Defines metrics to evaluate whether problems are due to infrastructure or application code, naming concrete observability tools and practices (logs, metrics… | ||
| FinOps on the Tutu IT Platform: How We Talk to the Business About Efficiency | Maksim Skomorokhov | 46m |
Overview of FinOps practices used on the Tutu IT platform to measure and improve cost efficiency: how the team quantifies platform costs, compares on‑premises… | ||
| Integrating Keycloak with Airflow, MLFlow, Superset and Monitoring Services | Aleksey Frolov | 35m |
Shows practical steps for integrating Keycloak SSO with ML infrastructure components: Airflow for orchestration, MLflow for model tracking, Superset for BI… | ||
| Geometry on Steroids: Practical Experience Porting Heavy Computation to GPU | Mikhail Lukin | 50m |
The talk covers using GPUs to accelerate computational geometry algorithms in CAD and engineering workflows, showing practical experience moving heavy… | ||
| CVE — Myth and Reality | Aleksey Morozov | 38m |
Covers several recent real CVEs for high-load systems, showing the lifecycle from researcher discovery to exploitation and demonstrating an example of… | ||
| Speeding Up Data Ingestion in PostgreSQL: From Custom Methods to Multithreading and Atomicity | D.Fatov | 45m |
Covers techniques to speed up data inserts into PostgreSQL, including batch INSERTs, custom insertion methods, and parallelizing the insert process. Explains… | ||
| Why There Aren't (Yet) Data Centers in Space | Aleksandr Tokarev | 48m |
The talk examines why full-scale data centers are not deployed in space, focusing on latency and bandwidth limits of current space communication channels. It… | ||
| Closing | Dasha Bykova | 13m |
Information about Saint HighLoad++ 2025, a professional conference for developers of high-load systems held on June 23–24, 2025 in Saint Petersburg at DESIGN… | ||
| Report on | 02m | |
A report about the Saint HighLoad++ professional conference for developers of high‑load systems held in Saint Petersburg (June 23–24, 2025) at DESIGN DISTRICT… | ||
| nothing matches | ||