Snowflake Administrator

Access that grows by exception becomes unmanageable. You build a role hierarchy with functional and access roles, use database roles and managed access schemas, and apply future grants so new objects arrive correctly protected.

• Roles, privileges and inheritance
• Functional roles and access roles
• Database roles and managed access schemas
• Future grants and ownership transfer
• Secondary roles and their trade-offs
• Auditing who has access to what

You build A role hierarchy with grant scripts for three teams and a review query.

Controlling who gets in is the first line of defense. You configure multi-factor authentication, single sign-on and automatic provisioning, key-pair access for services, and network policies and rules that limit where sign-in is allowed from.

• Multi-factor authentication and enforcement
• Single sign-on and SCIM provisioning
• Key-pair authentication for service accounts
• Network policies and network rules
• Private connectivity from your cloud
• Handling lockouts and emergency access

You build An authentication and network policy configuration for users and services.

Governance should follow the data. You classify sensitive columns, tag them, and attach masking and row access policies so protection applies the same way regardless of who runs the query.

• Classification of sensitive data
• Object tags and tag-based policies
• Dynamic data masking policies
• Row access policies
• Testing policies as different roles
• Reporting on policy coverage

You build A tag-driven masking and row access setup, tested as three different roles.

Day-to-day administration is largely about objects and their lifecycle. You manage databases, schemas and tables, set retention for Time Travel, clone for safe testing, and recover dropped objects.

• Databases, schemas and table types
• Time Travel retention and its storage cost
• Cloning objects for testing
• Undropping and restoring objects
• Storage metrics and cleaning up
• Managing changes to objects safely

You build A lifecycle plan with retention, cloning and recovery steps for one database.

Warehouses decide both speed and bill. You size and scale them for different teams, set auto-suspend and scaling policy, and use resource monitors to stop spend before it overshoots.

• Sizes, clusters and scaling policies
• Auto-suspend and auto-resume
• Separating teams and workloads
• Resource monitors and their actions
• Query timeouts and queuing limits
• Reviewing warehouse use over time

You build A warehouse layout for four teams with resource monitors and a monthly review query.

An administrator who learns of problems from users is late. You read account usage and information schema views, track queries, logins and credits, and set alerts that tell you first.

• ACCOUNT_USAGE against INFORMATION_SCHEMA
• Query, login and access history
• Credit and storage consumption
• Alerts and notifications
• Organization-level usage views
• Building an admin dashboard

You build A set of monitoring queries and alerts covering usage, cost and sign-in activity.

Users report slowness; you have to find the cause. You find costly queries, interpret spilling and queuing, and weigh clustering, search optimization, acceleration and warehouse changes against their cost.

• Finding the most expensive queries
• Reading profiles for spilling and queuing
• Clustering and search optimization choices
• Query acceleration and warehouse resizing
• Result caching and what invalidates it
• Telling users what to change in their queries

You build A tuning case file for a slow workload with the diagnosis and the change that fixed it.

Administrators often run the sharing process. You create and manage shares, set up reader accounts for outside consumers, publish and consume listings, and keep control of what recipients can see.

• Creating and managing outbound shares
• Consuming a share in another account
• Reader accounts and their administration
• Listings and the Marketplace
• Revoking access and auditing shares
• Sharing across regions

You build A share and a listing with controlled access, plus a review of who can see what.

Recovery plans only matter if they work. You replicate databases and account objects, use failover groups, understand backups and the limits of Time Travel and Fail-safe, and run a recovery exercise.

• Replication of databases and account objects
• Failover groups and promoting a secondary
• Backups and their retention
• Time Travel and Fail-safe limits
• Client connection redirect during failover
• Running a recovery exercise

You build A disaster recovery runbook with replication setup and a completed recovery exercise.

You set up and operate one account for a regulated team: roles, authentication, governance, warehouses, monitoring, sharing and recovery. Then you audit it against your own checklist and fix the gaps.

• Turning requirements into an account design
• Configuring access, governance and compute
• Setting up monitoring and sharing
• Rehearsing recovery
• Auditing the account and fixing gaps

You build A fully configured, documented account with an audit checklist and recovery runbook.