The article argues that as agents move from short, single‑shot tasks to long‑running workflows, reliability depends less on model correctness and more on clear state boundaries, evidence trails, and failure‑recovery loops that prevent erroneous submissions and make outcomes auditable.
AgentChord, a system presented at RSS 2026, anticipates potential robot manipulation failures by embedding recovery actions directly into a structured task graph, enabling immediate low‑latency switches to pre‑compiled recovery branches and achieving up to 99.2% success in simulated tasks and 77.5% on real robots.
This guide outlines core principles and practical solutions for building resilient backend systems, covering fast‑failure handling, automatic retries with exponential back‑off, circuit‑breaker usage, idempotency, batch job strategies, online transaction patterns, and robust message‑queue processing.
This article explains the Chandy‑Lamport algorithm for capturing consistent global snapshots in distributed systems, describes its assumptions and message‑marker rules, walks through a detailed example with three processes and channels, and relates it to Apache Flink's asynchronous checkpoint mechanism.
This article details the background, architecture, core features, and operational practices of a custom high‑reliability, high‑throughput key‑value storage system that combines Raft replication, flexible partitioning, binlog support, bulk loading, and multi‑active deployment for B‑Station's diverse data workloads.
Bilibili’s high‑performance distributed KV store combines hash and range partitioning, Raft‑based multi‑replica consistency, and a Metaserver‑managed topology of pools, zones, nodes, tables, shards and replicas, offering features such as partition splitting, binlog streaming, multi‑active replication, bulk loading, KV‑storage separation, and automated load, leader and health balancing for reliable, scalable data services.
This article describes a real‑world Ceph storage incident caused by disk bad sectors, analyzes its impact, and presents two practical recovery methods—full disk copy with dd+nc and skipping the faulty sector during service start—along with detailed commands and post‑recovery steps.
