Best of Google Cloud PlatformJune 2025

  1. 1
    Article
    Avatar of techleaddigestTech Lead Digest·45w

    When Google Sneezes, the Whole World Catches a Cold

    A detailed analysis of a major Google Cloud IAM outage that cascaded across multiple services including Cloudflare and Anthropic. The incident began with an IAM backend rollout issue at 10:50 AM PT, causing authentication failures across GCP products. Cloudflare's Workers KV, which depends on Google Cloud storage, failed next, affecting Access, WARP, and Zero Trust features. Anthropic disabled file uploads to manage error rates. Full recovery took over 7 hours, with some specialized services like Vertex AI requiring additional time. The analysis includes a detailed timeline, impact assessment, and lessons learned about dependency chains, control plane failures, and recovery patterns in distributed systems.

  2. 2
    Video
    Avatar of fireshipFireship·45w

    That time Google Cloud Platform bricked the Internet…

    Google Cloud Platform experienced a major outage that affected popular services like Snapchat, Spotify, Discord, and Cloudflare, causing nearly 100% error rates for over 2 hours. The incident was caused by a dormant bug in Google's API management service - a null pointer exception in code that lacked proper error handling. The bug was introduced on May 29th but remained undetected until a policy change on June 12th triggered the faulty code path, causing the API management binary to crash globally. Google took 40 minutes to begin rollback and 4 hours to fully stabilize, potentially costing millions in SLA credits and damaging their reputation in the competitive cloud market.

  3. 3
    Article
    Avatar of gcpGoogle Cloud·44w

    Tools Make an Agent: From Zero to Assistant with ADK

    Google's Agent Development Kit (ADK) enables building AI agents that can interact with external systems through various tool types. The framework supports function tools for inline calculations, built-in tools like Google Search, third-party API integrations via LangChain, and Model Context Protocol (MCP) tools for databases and APIs. A practical example demonstrates creating a bug assistant agent for a coffee company that can search tickets, query GitHub issues, and interact with PostgreSQL databases using these different tool types.