add README on GCP workload monitoring feature

MaxText/configs/base.yml

@@ -460,8 +460,8 @@ goodput_upload_interval_seconds: 60
 enable_pathways_goodput: False
 
 # GCP workload monitoring
-report_heartbeat_metric_for_gcp_monitoring: False
-report_performance_metric_for_gcp_monitoring: False
+report_heartbeat_metric_for_gcp_monitoring: True
+report_performance_metric_for_gcp_monitoring: True
 
 # Vertex AI Tensorboard Configurations - https://github.com/google/maxtext/tree/main/getting_started/Use_Vertex_AI_Tensorboard.md
 # Set to True for GCE, False if running via XPK

getting_started/GCP_Workload_Monitoring.md

@@ -0,0 +1,19 @@
+# Enable GCP Workload Monitoring
+This guide provides an overview on how to enable GCP workload monitoring for your MaxText workload.
+
+## Overview
+Google offers a monitoring and alerting feature that is well suited for critical MaxText workloads sensitive to infrastructure changes.
+Once enabled, metrics will be automatically sent to [Cloud Monarch](https://research.google/pubs/monarch-googles-planet-scale-in-memory-time-series-database/) for monitoring.
+If a metric hits its pre-defined threshold, the Google Cloud on-call team will be alerted to see if any action is needed. 
+
+The feature currently supports heartbeat and performance (training step time in seconds) metrics. In the near future, support for the goodput metric will also be added.
+Users should work with their Customer Engineer (CE) and the Google team to define appropriate thresholds for the performance metrics.
+
+This guide layouts how to enable the feature for your MaxText workload.
+
+## Enabling GCP Workload Monitoring
+User can control which metric they want to report via config:
+- To report the heartbeat metric, set `report_heartbeat_metric_for_gcp_monitoring` to `True`
+- To report the performance metric (training step time in seconds), set `report_performance_metric_for_gcp_monitoring` to `True`
+
+For an example, please refer to [base.yml](../MaxText/configs/base.yml).