Advanced AI Prompt for Service Mesh Traffic Management via GitHub Actions

Act as a Site Reliability Engineer. Design a GitHub Actions workflow that programmatically interfaces directly with the [SERVICE_MESH_TYPE] (e.g., Istio, Linkerd) control plane to orchestrate a highly granular traffic rollout. The workflow script must utilize `kubectl` or `helm` to sequentially patch the [VIRTUAL_SERVICE_NAME] CRD, mathematically incrementing the live traffic routed to the new container version by precisely [TRAFFIC_SPLIT_STEP]% every 2 minutes. Between every single iteration step, inject a verification sequence that queries the [GATEWAY_NAME] access logs or Prometheus endpoint for an increase in 5xx HTTP errors. If the error threshold is breached, the script must abort the loop and forcefully apply a hardcoded YAML reverting the VirtualService back to 100% stable routing.