Author

Random Bulletin

17-year internet software developer specializing in AI applications, networking, architecture, and open source. Led the delivery of network services handling hundreds of millions of concurrent devices and tens of millions of QPS, and has three years of experience designing and building an agent platform. Follow to stay updated.

115

Articles

Likes

311

Views

Comments

Latest from Random Bulletin

100 recent articles max

Random Bulletin

Aug 2, 2026 · Cloud Native

From Zero to Ten‑Million QPS: How to Reserve Resources for Guaranteed Capacity

The article explains how a noisy‑neighbor batch job can cripple a payment service at ten‑million‑QPS scale, then details five practical reservation techniques—quotas, reserved instances, isolation, priority preemption, and elastic prediction—while weighing their trade‑offs and showing how overcommit and offline mixing recover idle capacity for both high guarantee and high utilization.

Kubernetesovercommitpriority preemption

0 likes · 23 min read

From Zero to Ten‑Million QPS: How to Reserve Resources for Guaranteed Capacity

Random Bulletin

Aug 1, 2026 · Cloud Native

Mixed‑Tenant Architecture: From Exclusive to Shared, Boosting Utilization to 45%

The article explains how moving from exclusive, peak‑sized clusters to a mixed‑tenant architecture—leveraging time‑shifted workloads, cgroup and hardware isolation (LLC, memory bandwidth), QoS tiers, elastic throttling and dynamic over‑commit—can raise CPU utilization from under 20% to over 45% and cut costs by about 30%, while introducing significant operational complexity and stability risks.

CPU utilizationCloud NativeQoS

0 likes · 18 min read

Mixed‑Tenant Architecture: From Exclusive to Shared, Boosting Utilization to 45%

Random Bulletin

Jul 31, 2026 · Cloud Native

Scaling at Ten‑Million QPS: From Manual to Automatic Autoscaling

The article analyzes why manual capacity adjustments break down at ten‑million‑QPS scale, then walks through metric‑driven autoscaling, anti‑flapping algorithms, headroom planning, predictive scaling, stateful service challenges, and multi‑dimensional strategies to achieve a cost‑stable dynamic balance.

AutoscalingCloud NativeKubernetes

0 likes · 20 min read

Scaling at Ten‑Million QPS: From Manual to Automatic Autoscaling

Random Bulletin

Jul 30, 2026 · Operations

From Guesswork to Math: Modeling Capacity for Ten‑Million QPS Services

Capacity evaluation evolves from intuition‑based estimates to rigorous modeling by measuring per‑request costs with single‑machine benchmarks, applying USL and queueing theory, bounding effective capacity with SLOs, conducting full‑link shadow traffic tests, and continuously calibrating online water‑marks to turn high‑traffic resilience into a calculable engineering decision.

SLOUSLcapacity planning

0 likes · 19 min read

From Guesswork to Math: Modeling Capacity for Ten‑Million QPS Services

Random Bulletin

Jul 28, 2026 · Operations

From Wiki Docs to Executable Incident Plans: Cutting MTTR from Hours to Minutes

The article explains how to evolve static wiki‑based incident response plans into structured, executable, auditable systems—adding one‑click execution, gray‑scale rollbacks, permission controls, chaos‑engineered rehearsals, and alert integration—to reduce mean‑time‑to‑recovery from hours to minutes in high‑throughput environments.

MTTR reductionchaos engineeringincident response

0 likes · 19 min read

From Wiki Docs to Executable Incident Plans: Cutting MTTR from Hours to Minutes

Random Bulletin

Jul 27, 2026 · Backend Development

Isolation Strategies for Ten‑Million‑QPS Services: No Isolation, Thread‑Pool, Semaphore

The article explains how sharing a single thread pool can cause a slow downstream service to cripple an entire high‑traffic system, compares thread‑pool and semaphore isolation—detailing their strengths, costs, and suitable scenarios—and provides practical guidelines, framework support, and design decisions for implementing robust isolation in microservices.

ResilienceSemaphoreThread Pool

0 likes · 21 min read

Isolation Strategies for Ten‑Million‑QPS Services: No Isolation, Thread‑Pool, Semaphore

Random Bulletin

Jul 25, 2026 · Backend Development

Circuit Breaker Strategies: Evolving from Simple to Hierarchical for 10M QPS Systems

The article analyzes how a slow downstream service can trigger a site‑wide cascade failure, explains the three‑state circuit‑breaker model, compares simple global thresholds with layered multi‑metric, multi‑granularity approaches, and evaluates Hystrix, Sentinel, and Envoy implementations for high‑throughput back‑ends.

EnvoyHystrixSentinel

0 likes · 19 min read

Circuit Breaker Strategies: Evolving from Simple to Hierarchical for 10M QPS Systems

Random Bulletin

Jul 24, 2026 · Backend Development

Context Propagation at Ten‑Million QPS: From ThreadLocal to Explicit

The article examines a real‑world incident where ThreadLocal leakage caused user data mix‑ups, then walks through the evolution of context propagation—from implicit ThreadLocal to TTL patches, Reactor Context, ScopedValue, and Go's explicit context—highlighting trade‑offs and best practices for high‑throughput systems.

ContextPropagationGoContextReactor

0 likes · 15 min read

Context Propagation at Ten‑Million QPS: From ThreadLocal to Explicit

Random Bulletin

Jul 23, 2026 · Backend Development

From Threads to Coroutines: Scaling to Millions of QPS

The article explains why the traditional one‑request‑one‑thread model collapses at tens of thousands of connections due to memory, scheduling and cache costs, and shows how user‑space coroutine scheduling, stack‑less designs, Go's GMP model, Java virtual threads and CSP channels enable millions of concurrent connections with synchronous‑style code while highlighting trade‑offs and pitfalls.

CSPCoroutinesGo

0 likes · 21 min read

From Threads to Coroutines: Scaling to Millions of QPS

Random Bulletin

Jul 22, 2026 · Backend Development

Lock Optimization at Ten‑Million QPS: From Coarse Locks to Lock‑Free Designs

When a service’s QPS jumps from two to eight million, a single global lock can turn a multi‑core machine into a bottleneck, so the article walks through coarse‑lock problems, fine‑grained locking, read‑write separation, optimistic reads, CAS‑based lock‑free techniques, LongAdder and advanced data structures, showing when each step is appropriate for high‑throughput systems.

CASJavaLocks

0 likes · 19 min read

Lock Optimization at Ten‑Million QPS: From Coarse Locks to Lock‑Free Designs