You migrated applications to GCP but kept your existing monitoring; notifications are too slow for time-critical problems. What should you do?

The best approach is to leverage Stackdriver (now Cloud Monitoring and Cloud Logging) for its native integration with GCP, which provides real-time log capture and alerting capabilities. This directly addresses the slow notification issue for time-critical problems. You can then export these logs to your existing monitoring platform for historical analysis or consolidated dashboards, maintaining your current investment while improving responsiveness. Replacing your entire monitoring platform with Stackdriver is a significant undertaking and might not be necessary if your existing platform serves other needs well. Installing Stackdriver agents is good for collecting metrics and logs from VMs, but the core problem is slow notifications, which Stackdriver's alerting directly solves. Migrating traffic back is a workaround that doesn't solve the underlying monitoring latency.