Database Resilience, Observability & Modern Trends Quiz

Q1. [E] Which metric best indicates imminent WAL-archive saturation in a Postgres cluster? Checkpoint duration
Replication lag seconds
Archived WAL queue length
Dead-tuple count

Q2. [E] In a disaster-recovery plan, what does the Recovery Point Objective (RPO) specify for a database? The maximum period of data loss acceptable before harm
The longest allowable downtime for recovery
The size of the latest full backup
The number of standby replicas required

Q3. [M] To guarantee zero data loss in a failover event, which replication mode must a primary–standby database cluster use? Synchronous replication
Asynchronous replication
Single-master with nightly backups
Multi-master eventual consistency

Q4. [H] In a multi-node database cluster, what mechanism prevents a “split-brain” scenario (two primaries) during failover? Requiring a quorum/majority for leader election
Deploying exactly two nodes with auto-failover
Client-side load-balancing logic
Asynchronous replication to all replicas

Q5. [M] Which schema change is backward-compatible for zero-downtime deployments? Adding a new nullable or default-valued column
Renaming a frequently-used column in a table
Removing a column currently read by the app
Changing a column’s type from INT to DATETIME

Q6. [M] A consistently low buffer-cache hit ratio in a database indicates that: The server is often reading data from disk rather than memory
Queries are perfectly optimized and use the cache fully
The database has excess memory not being utilized
CPU, not I/O, is the primary performance bottleneck

Q7. [E] Which feature ensures that database data remains encrypted at rest (on disk)? Transparent Data Encryption (TDE)
Parameterized SQL queries
Role-based access control
TLS (SSL) connection encryption

Q8. [E] When investigating a slow SQL query, what is typically the first step for diagnosing the performance issue? Examine the query’s execution plan (EXPLAIN)
Add indexes on every column in the query
Restart the database server to clear caches
Denormalize the related database tables

Q9. [M] Which statement is true of serverless database services? They auto-scale resources per demand and bill per usage, with no user-managed servers
They run without using any physical servers or hardware whatsoever
They require manual capacity planning like traditional databases
They are limited to development use and cannot handle production loads

Q10. [H] What is a primary characteristic of a Hybrid Transaction/Analytical Processing (HTAP) database? It handles OLTP (transactions) and OLAP (analytics) on the same real-time data simultaneously
It offloads transactions to an OLTP database and analytics to a separate OLAP warehouse
It uses GPUs to accelerate transactions
It performs analytics only on stale read replicas of the primary database

Q11. [M] What are vector databases optimized for? Storing high-dimensional embedding vectors and performing similarity searches on them
Managing time-series numeric data with time-based queries
Storing GIS spatial coordinates and performing geometric queries
Holding relational tables with strict schemas and foreign-key enforcement

Q12. [H] Which emerging I/O technology can drastically reduce write latency by providing persistent storage at near-DRAM speeds? Non-volatile memory (persistent memory like NVDIMM/Optane)
15k RPM enterprise HDDs
Standard SATA SSD drives
Dual 10-Gigabit network cards

Q13. [M] Adding an index on a frequently-filtered column will typically: Speed up SELECT queries on that column, but slightly slow down writes and use more storage
Speed up all queries (reads and writes) with no downsides
Make insert/update operations faster while making reads slower
Eliminate the need for ACID transactions on that table entirely

Q14. [H] A database experiencing high lock contention is likely suffering from: Many concurrent transactions trying to update the same rows simultaneously
Too many SELECT (read-only) queries running in parallel
The use of connection pooling across the application
Excessive checkpointing or logging by the database engine

Q15. [E] Applying the “principle of least privilege” in database security means: Granting each user/service only the minimum permissions needed to perform its tasks
Using one shared admin account for all application and user access to the DB
Revoking all privileges from every user as a default security posture
Giving developers full administrative access in non-production environments

system-design