Network Performance Dashboard shows increased latency and loss to a partner site in another region. To determine whether the issue occurs inside Google's network or after egress to the public internet, what is the recommended next step?

Running an on-host traceroute from an affected VM is the most direct way to diagnose where latency and loss are introduced. Traceroute reveals the hop-by-hop path, allowing you to see if performance degrades within Google's network (internal IP addresses) or after traffic exits to the public internet (external IP addresses). This helps pinpoint whether the issue is with Google's infrastructure or an external ISP. Connectivity Tests are useful for verifying reachability and firewall rules but don't provide the granular hop-by-hop latency details needed for this specific diagnosis. Packet Mirroring captures traffic but doesn't inherently attribute retransmissions to a specific network segment without further analysis. Scaling up the VM's network interface addresses host-side bottlenecks, which is a different diagnostic path than identifying network path issues.