BestHub
Sign in
Tagged articles

overcommit

10 articles · Page 1 of 1
Random Bulletin
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
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 utilizationQoSResource Isolation
0 likes · 18 min read
Mixed‑Tenant Architecture: From Exclusive to Shared, Boosting Utilization to 45%
Didi Tech
Didi Tech
Oct 19, 2023 · Cloud Native

Design and Implementation of a New Tiered Resource Guarantee System for Elastic Cloud Containers

The new tiered resource‑guarantee system for Didi’s elastic cloud containers defines S, A, and B priority levels with explicit over‑commit rules, upgrades OS, Kubernetes, kube‑odin, service‑tree, and CMP components, and thereby cuts CPU contention by up to 80%, reduces latency, improves scaling reliability, and lowers operational costs.

Container ManagementKubernetesQuota
0 likes · 16 min read
Design and Implementation of a New Tiered Resource Guarantee System for Elastic Cloud Containers
High Availability Architecture
High Availability Architecture
May 26, 2023 · Big Data

Amiya: Dynamic Overcommit Component for Bilibili Offline Big Data Cluster Resource Scheduling

This article introduces Amiya, a self‑developed overcommit component that dynamically increases Yarn memory and vCore capacity on Bilibili's offline big‑data clusters, details its architecture, key implementation of overcommit, eviction and mixed‑deployment strategies, and evaluates its resource‑utilization impact.

Resource OptimizationYARNcluster-management
0 likes · 22 min read
Amiya: Dynamic Overcommit Component for Bilibili Offline Big Data Cluster Resource Scheduling