Cloud Network Insights Cloud Network Insights: end-to-end observability for the Cross-Cloud Network

Summary

Poonam Yadav, a product manager within Google Cloud Networking, introduces Cloud Network Insights, a fully managed observability product designed to provide absolute clarity for distributed, multi-cloud, and hybrid network performance. Modern application architectures increasingly involve agentic workflows distributed across multiple clouds, on-premises data centers, SaaS services, internet endpoints, and transit networks. Because customers use multiple clouds like Google Cloud, Azure, and AWS, along with managed services like MongoDB and Databricks, network architectures have become highly complex. Cloud Network Insights addresses this complexity by offering end-to-end network performance visibility powered by artificial intelligence.

The product monitors various components including Google Cloud, third-party clouds, on-premises data centers, collocations, internet-facing services, ISPs, transit networks, interconnects, and third-party APIs and agents. It tracks network availability, latency, loss, jitter, round-trip time, web application performance, and emerging agentic protocols. By integrating deeply with Google Cloud intelligence, AI, metrics, and logs, Cloud Network Insights helps operators reduce mean time to innocence and mean time to remediation. It features natural language querying through Gemini Cloud Assist and supports network performance monitoring for agentic workloads.

Cloud Network Insights works through monitoring points, which act as probing agents installed in virtual machines or containers anywhere in the network. These monitoring points send synthetic traffic across monitored endpoints and collect performance metrics, sending them to a backend SaaS service and making them available in the Google Cloud console. Network monitoring operates at layers three and four of the OSI model using continuous lightweight traffic, TCP, ICMP, and UDP path monitoring, and packet train dispersion technology. Web monitoring covers layer seven, providing comprehensive HTTP and synthetic transaction testing, scripting milestones for single-page apps, and multi-page waterfall charting.

During the product demonstration within the Google Cloud console, the interface displays monitoring points installed in locations such as Taipei, US Central, Azure, and on-premises in Tel Aviv. Users can configure network paths and web paths to track metrics such as average and maximum round-trip time, inbound and outbound data loss, and jitter. The demonstration highlights the ability to troubleshoot paths using a troubleshooting UI that shows hop-by-hop routes and route change events during data loss incidents. For web monitoring, the demo illustrates HTTP server response breakdowns, resource timing, download speed, and browser workflow testing with milestone breakdowns and end-user experience tracking across network, server, and browser components.

The pricing model is designed to be disruptive, flexible, and competitive based on the number of active paths being monitored rather than the number of tests run over those paths. Three licensing options are available. The pay-as-you-go option suits customers trialing the product or needing monitoring for only a few days a month. The enterprise option provides a cost-effective choice for continuous monitoring on a set number of paths. The unlimited option uses an attractive flat annual fee for large-scale deployments covering thousands of paths. Additionally, Cloud Network Insights integrates with Google Cloud partners and supports the OpenTelemetry Protocol for direct ingestion of metrics and logs into tools like Datadog, Dynatrace, New Relic, Grafana Labs, PagerDuty, XMatters, and ServiceNow.