Get the weekly digestSubscribe

/ THE FEED /

Read what ships.

Deep, recent reads ranked by depth and freshness.

Distributed Systemsremove filterReliabilityremove filterClear all

  1. 01The Pragmatic Engineer · architecture

    The Pulse: Anthropic builds its own source control system

    Anthropic experienced a massive increase in code merges due to the heavy use of AI coding agents, which quickly overwhelmed standard Git scaling limits and CI checkouts. To solve this, a small engineering team built a custom centralized version control system called Scion backed by Google Spanner and Cloud Storage, along with a code review tool called Antique. This move successfully migrated the entire monorepo off Git and eliminated bottlenecks, though the architecture drew controversy over similarities to Meta's Mononoke.

    DEPTH ■■■■■12 min read4 days ago

    ArchitectureDistributed Systems

    Sign in to save and rate

  2. 02Grab · architecture

    App stability at scale with a layered defense

    To maintain app stability amidst rapid weekly releases and AI-generated code, Grab uses a layered defense strategy covering prevention, release gates, and incident response. This pipeline includes static analysis via custom lint rules, multiple tiers of automated and manual testing, a pre-production early access build, staged rollouts, and feature flags with instant kill switches. As a result, critical defects are caught early in the development lifecycle before reaching the wider production user base.

    DEPTH ■■■■■16 min read5 days ago

    MobileReliability

    Sign in to save and rate

  3. 03Zalando · postmortem

    Thundering herd of circuit breakers

    An engineering team investigated why a client-side circuit breaker repeatedly failed to close on its own after an outage. By exporting OpenTelemetry traces into a custom visualization tool, they discovered that low probe traffic caused the autoscaler to scale down the service to minimum capacity, triggering a thundering herd of requests and immediate timeouts the moment the circuit breaker closed. Adjusting the system to account for capacity and probe representativeness prevents this oscillation.

    DEPTH ■■■■■10 min read2 weeks ago

    MicroservicesObservabilityReliability

    Sign in to save and rate