Security Setup Checklist
70 steps across the ten Brilliant at the Basics requirements. Check each box as you complete it. Confirm every step against Google Cloud’s own documentation before you rely on it.
For CUI, deploy inside Assured Workloads with the correct US government / Impact Level compliance regime.
Phishing-Resistant Multi-Factor Authentication
- Open the Admin consoleGoogle Cloud sign-in is managed in Google Workspace or Cloud Identity. Go to admin.google.com and sign in as an administrator.
- Find 2-Step Verification settingsIn the Google Admin console, go to Security > Authentication > 2-Step Verification.
- Turn on enforcementCheck 'Allow users to turn on 2-Step Verification' and set enforcement to On, choosing an enrollment grace period from 1 day to 6 months.
- Require security keys onlyUnder 'Methods', pick 'Only security key' so only hardware keys and passkeys are allowed. This blocks weaker codes that phishing can steal.
- Buy and register keysGive each admin a FIDO2 hardware key such as a Titan Security Key, then have them register it at myaccount.google.com under Security > 2-Step Verification.
- Protect admin accounts firstApply enforcement to an organizational unit or group that contains your administrators and users who touch sensitive data.
- Set backup keysHave each user register a second key so a lost key doesn't lock them out; store the spare in a safe place.
Evidence to keep: Screenshot of 2-Step Verification enforcement setting · List of users with security keys enrolled · Admin console audit log of the policy change
Least-Privilege Access Control
- Open IAMIn the Google Cloud console, go to IAM & Admin > IAM to see every person and service account and the roles they hold.
- Use predefined rolesWhen granting access, click 'Grant Access' and pick a narrow predefined role (like 'Storage Object Viewer') instead of broad roles like Owner or Editor.
- Check role recommendationsLook for the light-bulb recommendation icon next to a principal in IAM, or go to IAM & Admin > Recommendations to see where IAM Recommender suggests smaller roles.
- Apply the tighter roleClick the recommendation and apply it to swap an over-broad role for one that still covers what the user actually does.
- Create dedicated service accountsIn IAM & Admin > Service Accounts, make a separate service account for each app or workload so programs don't share one powerful identity.
- Lock down key creationIn IAM & Admin > Organization Policies, enforce the constraint 'Disable service account key creation' to stop risky downloadable keys.
- Review access regularlySet a monthly reminder to revisit IAM and Recommendations and remove any access that is no longer needed.
Evidence to keep: Export of IAM policy bindings per project · IAM Recommender history showing applied changes · Organization policy report on service account key constraints
Asset & Account Inventory
- Open Asset InventoryIn the Google Cloud console, go to IAM & Admin > Asset Inventory to see all resources across your project, folder, or organization.
- Pick your scopeUse the scope selector at the top to view assets at the organization, folder, or project level so you see the full picture.
- Browse resource typesUse the Resources tab and filters to list VMs, buckets, and databases; click a resource's name to view its metadata.
- Review the hierarchyGo to IAM & Admin > Manage Resources (Resource Manager) to see the tree of organization, folders, and projects that groups everything.
- Export a snapshotUse the Export option in Asset Inventory (or the gcloud CLI) to save a point-in-time list to BigQuery or Cloud Storage for your records.
- Enable change trackingTurn on Asset Inventory feeds so you get notified when resources are created, changed, or deleted.
- List human and service accountsIn IAM & Admin > IAM and > Service Accounts, record who and what has access so people are inventoried too.
Evidence to keep: Exported asset list from Asset Inventory · Resource Manager hierarchy diagram or screenshot · List of all user and service accounts
Logging, Monitoring & Audit
- Open Logs ExplorerIn the Google Cloud console, go to Logging > Logs Explorer to search your logs; write a query and click Run query.
- Turn on Data Access logsGo to IAM & Admin > Audit Logs, pick the services you care about, and enable DATA_READ and DATA_WRITE logging (these are off by default).
- Confirm Admin Activity logsAdmin Activity audit logs are always on; verify you can see them in Logs Explorer so account and config changes are tracked.
- Route logs to a safe sinkIn Logging > Log Router, create a sink that sends logs to a locked Cloud Storage bucket or BigQuery for long-term retention.
- Activate Security Command CenterIn the console, go to Security > Security Command Center and activate it at the organization level to surface threats and misconfigurations.
- Set alertsIn Monitoring > Alerting, create alerting policies (for example on failed logins or IAM changes) so someone is notified fast.
- Protect the logsRestrict who can edit or delete log sinks and buckets so an attacker can't erase their own tracks.
Evidence to keep: Screenshot of enabled Data Access audit log config · Log Router sink pointing to retained storage · Security Command Center findings report
Network Segmentation & Boundary Protection
- Create a VPC networkIn the Google Cloud console, go to VPC network > VPC networks and click Create VPC network to build your private network.
- Add tiered subnetsWithin the VPC, create separate subnets for each tier (web, app, database) so functions are isolated by design.
- Write firewall rulesGo to VPC network > Firewall and create rules that deny by default and allow only needed traffic, targeting instances by network tags or service accounts.
- Set up Shared VPCIn VPC network > Shared VPC, designate a host project and attach service projects for centralized, consistent network control.
- Build a service perimeterGo to Security > VPC Service Controls, click Manage access policies, and create a perimeter around projects so data can't leak to outside services.
- Add host and service projectsWhen using Shared VPC, add both the host project and the service projects to the perimeter so the whole boundary is covered.
- Test then enforceRun the perimeter in dry-run mode first to see what it would block, then switch to enforced once it looks correct.
Evidence to keep: VPC and subnet configuration export · Firewall rule list showing default-deny · VPC Service Controls perimeter configuration
Vulnerability & Patch Management
- Enable VM ManagerIn the Google Cloud console, go to Compute Engine > VM Manager and enable the OS Config service so agents can inspect your VMs.
- Turn on OS inventoryEnsure the OS Config agent is installed and OS inventory is enabled so each VM reports its installed packages.
- Review vulnerability reportsIn Compute Engine > VM Manager > Vulnerability reports, view the CVEs (Common Vulnerabilities and Exposures) found on your machines.
- Create patch jobsIn VM Manager > Patch, create a patch deployment to update Windows and Linux VMs on a schedule instead of by hand.
- Scan container imagesTurn on Artifact Analysis in Artifact Registry so container images are automatically scanned for known vulnerabilities.
- See it all in SCCGo to Security > Security Command Center to view VM and container findings across all projects in one dashboard.
- Prioritize and fixSort findings by severity in SCC and patch the critical and high items first, tracking each to closure.
Evidence to keep: VM Manager vulnerability report export · Patch deployment job history · Security Command Center findings list with remediation status
Data Protection & Encryption
- Open Key ManagementIn the Google Cloud console, go to Security > Key Management to manage your encryption keys.
- Create a key ringClick Create key ring, give it a name, and choose a location that matches the region of the data you'll protect.
- Create a keyInside the key ring, click Create key to make a CMEK, and set a rotation period so the key changes automatically over time.
- Grant the service accountGive the resource's service account the role 'Cloud KMS CryptoKey Encrypter/Decrypter' on that key so it can use it.
- Attach CMEK to resourcesWhen creating a bucket, disk, or database, choose 'Customer-managed encryption key' and select your CMEK.
- Confirm default encryptionRemember Google already encrypts all data at rest by default; CMEK adds your own control on top for sensitive workloads.
- Enforce encryption in transitRequire TLS (HTTPS) for connections and enable 'Enforce SSL' options so data is protected while moving between systems.
Evidence to keep: Cloud KMS key ring and key configuration · Resource settings showing CMEK enabled · Key rotation schedule record
Backup & Recovery
- Open Backup and DRIn the Google Cloud console, go to Backup and DR to reach the data protection interface.
- Create a backup vaultCreate a backup vault as an isolated, protected storage destination, and choose a region or multi-region location for resilience.
- Define a backup planCreate a backup plan setting the frequency, retention period, and the backup vault where copies are stored.
- Protect your VMs and disksApply the backup plan to your Compute Engine instances and Persistent Disks (or use the default backup plan) so they are captured automatically.
- Add snapshot schedulesIn Compute Engine > Snapshots, attach a snapshot schedule to disks for simple, low-cost automated backups into your chosen region.
- Store copies in another regionUse multi-region vaults or multi-region snapshot storage so a single region outage doesn't take your backups with it.
- Test a restorePeriodically restore a backup into a test environment to prove the copies actually work before you need them.
Evidence to keep: Backup plan and vault configuration · Snapshot schedule showing recent successful backups · Documented restore test results
Secure Configuration Baseline
- Activate SCC at org levelIn the Google Cloud console, go to Security > Security Command Center and activate it at the organization level to enable the security posture service.
- Apply a CIS posture templateIn Security Command Center > Posture, deploy the predefined CIS Google Cloud Benchmark posture template to detect settings that don't meet the standard.
- Set organization policiesGo to IAM & Admin > Organization Policies and enforce constraints (like disabling public IPs or service account keys) to prevent risky configurations.
- Review Security Health AnalyticsIn Security Command Center, check the Security Health Analytics findings for misconfigurations across your projects.
- Deploy postures by scopeApply postures at the organization, folder, or project level so the right baseline covers the right resources.
- Fix driftWork through the findings and correct each resource that has drifted from the baseline, then re-check that it clears.
- Monitor continuouslyKeep the posture active so new violations are flagged automatically instead of only at audit time.
Evidence to keep: Security posture definition and compliance report · Organization policy constraint settings · Security Health Analytics findings showing resolved items
Secure AI Adoption & Data Loss Prevention
- Open Sensitive Data ProtectionIn the Google Cloud console, go to Security > Sensitive Data Protection to reach the scanning and de-identification tools.
- Run a discovery scanCreate an inspection or discovery job to scan Cloud Storage and BigQuery for sensitive data using the built-in detectors (150+ types).
- De-identify sensitive fieldsUse de-identification templates to mask or tokenize sensitive values before data is used to train or prompt AI models.
- Secure Vertex AI accessIn Vertex AI, restrict access with IAM roles and put Vertex AI resources inside a VPC Service Controls perimeter to stop data leaving.
- Filter AI inputs and outputsRoute prompts and responses through Sensitive Data Protection (or Model Armor) so sensitive text is caught in both directions.
- Turn on audit loggingEnable Cloud Audit Logs for Vertex AI so every model and data access is recorded for review.
- Set an AI use policyWrite a simple rule for staff on what data may and may not be sent to AI tools, and train the team on it.
Evidence to keep: Sensitive Data Protection scan results and de-id templates · VPC Service Controls perimeter covering Vertex AI · Vertex AI audit logs and written AI use policy