When the Cloud Controller Blobstore is set to Google Cloud Platform (GCP) Cloud Store, you are experiencing a sharp rise in the time it takes to complete a cf push command.
This issue is related an overburdened Cloud Controller or Cloud Controller Worker. You can use the cf firehose plugin and run the following commands to see if this is the case:
// check CPU load: if value is greater then 1, your CC is overloaded (scale horizontally) cf nozzle -no-filter | grep "origin:\"cc\"" | grep cpu_load_avg // check requests outstanding: If greater then 20, CC or CC workers needed to be scaled cf nozzle -no-filter | grep "origin:\"cc\"" | grep "requests.outstanding"
If the metrics above seem to be normal, then it is possible that the GCP blobstore is causing delays with Resource Matching.
The following screenshot shows how Healthwatch is showing high push times and when these times trigger an alert.
cf curl -X PATCH /v3/feature_flags/resource_matching -d -H "Content-type: application/json" -d '{"enabled": false}'