Amazon S3

Evidence (9)

• 8Hacker News·8/12/2026open_source
  Tailscale Traces Database Corruption to a 16-Year-Old SQLite WAL Bug

  Tailscale says months of repeated database corruption incidents in its control plane were ultimately traced to a long-standing SQLite WAL-reset bug. The company reports it fixed the issue and is publishing the incident details after 19 corruption events over six months that caused shard-level downtime and some lost configuration changes.

• 6InfoQ·8/10/2026framework_update
  Canva Uses S3 to Scale Session Revocation

  Canva redesigned its session revocation system to handle around 100 million active sessions while reducing database lookups. The new approach stores durable revocation records in Amazon S3 and distributes compact in-memory indexes to application gateways, cutting revocation cache memory usage by 87.5% and simplifying deployment.

• 5InfoQ·7/30/2026framework_update
  AWS Lambda Raises Managed Code Storage Quota

  AWS Lambda now lets customers reference deployment packages directly from their own S3 buckets, which removes the per-Region managed code storage quota and increases the default managed limit from 75 GB to 300 GB. Per-function package size limits remain unchanged, and users still need to run UpdateFunctionCode after replacing an object.

• 6The New Stack·7/23/2026research
  Prompt Caching and Batched Pipelines to Cut RAG Costs

  The article argues that naive retrieval-augmented generation (RAG) architectures often fail in production because synchronous ingestion, timeout limits, and rate limits do not scale to enterprise workloads. It recommends asynchronous uploads, micro-batching, and controlled embedding workers, and discusses prompt caching as one way to reduce costs without hurting accuracy.

• 6The New Stack·7/15/2026product_launch
  What breaks when AI notetaker memory scales

  The article examines how Plaud’s AI notetaker hit a scaling problem not in transcription quality, but in serving stored transcripts back to users quickly. Its original split between MySQL for metadata and S3 for large audio and transcript files worked at launch, but created latency and consistency issues when the product needed to behave like a real-time memory system.

• 6The New Stack·7/9/2026research
  Autonomous ingestion errors poison a vector database

  A fintech team describes how an autonomous PDF ingestion pipeline inserted incorrect metadata into a vector database, causing a chatbot to cite outdated and wrong financial data. The post-mortem says the root cause was treating probabilistic LLM extraction as deterministic, and that a second LLM-based validation step failed due to mutual confirmation bias.

• 6InfoQ·7/8/2026framework_update
  Airbnb Details Sitar-agent for Dynamic Kubernetes Configuration

  Airbnb engineers described Sitar-agent, a Kubernetes sidecar used to deliver dynamic configuration to tens of thousands of pods with updates arriving several times per minute. The system was redesigned around Java, Amazon S3 snapshot bootstrapping, and a migration from Sparkey to SQLite to improve reliability, startup speed, and configuration availability at scale.

• 6InfoQ·7/5/2026product_launch
  AWS Adds Amazon S3 Annotations

  AWS has introduced Amazon S3 Annotations, a feature for attaching searchable context such as summaries, classifications, compliance data, and AI-generated insights to S3 objects. The annotations can be updated independently of the underlying object and queried across datasets, which may reduce reliance on separate metadata systems.

• 7Hacker News·7/2/2026product_launch
  ZeroFS introduces a log-structured filesystem for S3

  ZeroFS is an open-source filesystem that serves S3-compatible storage as POSIX filesystems over NFS and 9P, and as raw block devices over NBD. It writes data as immutable, compressed, encrypted segments and includes local caching, checkpoints, replication, and durability semantics for different mount types.