Ouch. What resources can you not remove? What exactly are you running?
In general, as first thing stop the bleeding:
1. Stop your services from running
2. Check your IAM policies for anything suspicious, new service accounts, new users. Clean up.
3. Rotate all your Service Accounts and Service Account’s keys! If possible re-provision your machines (with a new SA) and redeploy your apps.
4. Check your VPC’s firewall
Then you absolutely need to figure out how you’ve been hacked. If the breach is on the application layer you must figure out where and patch it. Check your application logs.
Then check your GCP activity logs, search for unexpected calls from service accounts - assume the attacker has compromised a service account and search for attempt to persist with calls to `setIam` or other sensitive api calls.
Sorry, I’m on mobile but feel free to reach out If you need (email in profile)