BRIEF #2
February 1, 2026

Platform Pulse: Sub-100ms Serverless Cold Starts, SQL-Native AI, and the Bangkok Region Launch

This edition covers a breakthrough in Cloud Run cold start performance, the integration of open models directly into BigQuery SQL, and the official launch of the new asia-southeast3 region. Plus, critical updates on Firebase relational support and Service Mesh patching.

⚡ Community Corner

Best articles from the wider community and deep-dive tutorials.

  1. Google Found a Way to Make Serverless Actually Fast: A look at how Cloud Run has significantly improved serverless function performance to achieve sub-100ms cold starts. This optimization utilizes a combination of startup CPU boosts, lightweight gVisor sandboxing, and efficient concurrency management.

  2. Firebase Just Got Relational: SQL in Next.js 16: For developers modernizing frontend applications, this guide introduces how Firebase Data Connect now allows integration with relational databases (PostgreSQL via Cloud SQL) directly into Firebase projects.

  3. Deploying the Neo4j MCP Server to Cloud Run: As we move toward agentic architectures, this tutorial covers strategies for deploying the Neo4j Model Context Protocol (MCP) server on Cloud Run for both development and production environments.

  4. How Google SREs Use Gemini CLI to Solve Outages: A practical look at how Google's own SREs utilize the Gemini CLI and Gemini 3 models to automate incident response, effectively reducing "Bad Customer Minutes" and eliminating toil.

  5. BigQuery Cost Spikes Explained: A FinOps essential reading that explains how minor query changes can inadvertently increase bytes processed or shuffled, triggering 10x cost spikes, and how to prevent them.

🚀 Official Release Notes & Updates

Official product announcements and feature releases.

  1. New Region: Bangkok (asia-southeast3): Google Cloud has officially launched the asia-southeast3 region in Bangkok, Thailand to support low-latency and PDPA compliance. Services immediately available include Compute Engine, GKE, Cloud Storage, BigQuery, AlloyDB, and Artifact Registry.

  2. Cloud Run & Functions Python Support: Source deployments for Cloud Run and Cloud Run functions now support the pyproject.toml file for managing dependencies in General Availability. Additionally, the Python buildpack now supports default entrypoint detection for the Agent Development Kit (ADK) framework.

  3. BigQuery Managed Inference for Open Models: BigQuery now supports managed and SQL-native inference for open models from Hugging Face and Vertex AI Model Garden. You can now specify which model to use by passing an endpoint argument to functions like AI.IF, AI.SCORE, and AI.CLASSIFY.

  4. Cloud SQL for MySQL Performance Capture: Now in Preview, this feature helps diagnose transient performance issues by taking snapshots of the database state (including in-flight transactions and InnoDB status) when configurable thresholds are reached.

  5. Firestore Enterprise Native Mode Updates: Firestore Enterprise edition now supports Native mode and the Pipeline operations interface, adding complex expressions and functions like array_contains_all and regex_match.

  6. Load Balancing Traffic Duration & Backend Buckets: Application Load Balancers now support configuring traffic duration (SHORT or LONG) based on backend response needs. Additionally, backend buckets are now available for regional external and internal Application Load Balancers.

🛡️ Security & Advisories

CVEs, Bulletins, and Critical Patches.

  1. Managed Workload Identity for Load Balancers: Now in Preview for global external Application Load Balancers, this feature enables automated certificate rotation and trust management for backend mTLS via integration with Certificate Authority Service.

  2. Chronicle Ingestion for Model Armor: You can now ingest Google Cloud Model Armor logs (GCP_MODEL_ARMOR) directly into Google SecOps to monitor AI-specific threats like prompt injection and sensitive data leakage.

  3. DNS Armor General Availability: Monitoring internet-bound DNS queries for malicious activity using DNS Armor is now generally available (GA).

  4. Mandiant Net-NTLMv1 Deprecation Tools: To accelerate the deprecation of the insecure Net-NTLMv1 protocol, Mandiant has released rainbow tables to help security professionals demonstrate its insecurity.

  5. Cloud Service Mesh Patching: New patch releases (1.21.6-asm.8, 1.20.8-asm.60, 1.19.10-asm.55) are rolling out to address CVE-2025-61729 and CVE-2025-61727 in the Proxy and Control Plane.

0

From the Community

No community links this week.

Enjoyed this brief?

Don't miss the next drop.