Databricks Platform Admin cheat sheet
Databricks Platform Admin · Keep the key ideas within reach.
Databricks accounts, workspaces and identity boundaries
Account, workspace, data-governance and compute boundaries differ.
Provisioning, sign-on, workspace entitlement and data authorization are separate.
Automated workloads need an intentional Run as identity.
Current cloud-specific identity and networking features need actual configuration evidence.
Unity Catalog objects, privileges and storage responsibility
Resolve data privileges together with parent usage and binding conditions.
Storage credentials, locations and table metadata have distinct responsibilities.
Revoking one grant does not prove all effective access disappeared.
Tags, lineage and sharing need actual enforcement and capture context.
Databricks compute policies, jobs and operational evidence
Policy constraints, policy use rights and data access are distinct.
A job’s actual graph and Run as identity determine its behavior.
Concurrent-run limits need queue settings and trigger context.
Audit, query, task and transaction histories answer different questions.
SQL models, types and constraints
SQL tables may contain duplicate rows unless keys or other constraints prevent them.
PRIMARY KEY is unique and nonnull; CHECK accepts TRUE or UNKNOWN.
Choose exact numerics for exact decimal requirements.
DDL transaction and identity effects depend on the specified DBMS.
SQL retrieval, joins and transaction semantics
WHERE retains TRUE predicates; NULL comparisons can be UNKNOWN.
Keep optional-side filtering consistent with the intended outer-join result.
Use a guaranteed nonnull marker for an unmatched-row test.
Check join cardinalities before trusting an aggregate total.
Identity, access and cryptographic protection
Identification, authentication, authorisation and accounting are four distinct steps.
MFA needs two different factor categories, not two instances of one.
RBAC suits stable job roles; ABAC suits context-dependent, fine-grained rules.
Kerberos uses a KDC with TGT and service tickets for SSO.
Symmetric ciphers handle bulk data; asymmetric ciphers solve key distribution and signatures.
Signing uses a specified signature scheme with the signer’s private key; verification and custody are essential to origin evidence.
ECB leaks patterns; use authenticated modes such as GCM or encrypt-then-MAC.
Network boundaries and protected communication
Internet-facing services belong in a DMZ; data stores sit in a filtered internal zone.
VLANs alone are not a trust boundary - add DAI, private VLANs and 802.1 X.
IPsec AH gives integrity only; ESP adds confidentiality.
IDS is passive and alerts; IPS is inline and blocks.
NAT is not a security control and temporary rules must be removed.
Security operations, monitoring and response
Asset inventory and configuration baselines underpin every operational control.
Changes need an appropriate authorization path, recovery planning and relevant evidence.
Emergency changes still require retrospective documentation within a defined window.
Monitor log sources for silence, not just for events.
Distinguish historical Revision 2 phases from current NIST incident-response guidance integrated with broader risk management.
Containment precedes eradication in almost every scenario.
Capture volatile evidence (memory, connections) before disk and before power-off.
Continuity, backup and technical recovery
Distinguish tolerated data loss from tolerated interruption.
A recovery point and restore time must both meet the stated targets.
A same-failure-domain snapshot alone does not cover loss of that domain.
Verify complete restore procedures and actual dependency behavior.