Interview
Say the answer out loud
Panel questions and concept drills. Not the engine course, that is Lessons. Come here to rehearse on the same evidence.
Need the mechanism first? Open Lessons. Want a study order? Pathways.
Start here
2 free panel questions, full worked answers
- Free · midHow does VACUUM relate to MVCC, and what happens if it falls behind?Someone who can define MVCC and VACUUM separately but needs to connect them convincingly in an interview.
- Free · midYou want to measure a table's access pattern cleanly after a change. How do you reset just that table's statistics without wiping the cluster?Someone who needs to clear one table's activity counters to measure a change, and to explain why that is not the same as ANALYZE.
Prefer one mechanism at a time? Open concept drills · Need the engine course first? Lessons
No fabrication. Every technical answer that cites a number was reproduced on a real PostgreSQL 18 lab, and the exact transcript ships with the Pro answer so you can check it. The behavioural prompts coach you on structure only. They do not invent a story for you, because an interviewer catches a borrowed one in the very next follow-up.
Showing 19 of 19
- seniorCore PostgreSQLPro answer
Explain REPEATABLE READ vs SERIALIZABLE in PostgreSQL and when you would choose each.
Someone preparing to explain, out loud, why they would pick one PostgreSQL isolation level over the other.
Open the question → - midCore PostgreSQLFree in full
How does VACUUM relate to MVCC, and what happens if it falls behind?
Someone who can define MVCC and VACUUM separately but needs to connect them convincingly in an interview.
Open the question → - staffCloud Postgres platformPro answer
You've got a 50 TB table. Walk me through what ANALYZE actually does at that size, does it read every page?
Someone who needs to explain ANALYZE's sampling behaviour on a very large table without overclaiming that it is cheap.
Open the question → - seniorCloud Postgres platformPro answer
Take me from that sample to a query plan, what does ANALYZE compute, and how does the planner use it?
Someone who needs to explain how PostgreSQL turns column statistics into a row estimate, and what to do when the estimate is wrong.
Open the question → - staffCloud Postgres platformPro answer
Give me a multixact scenario and take it deep, when does one get created, and how does it bite you in production?
Someone asked to give a concrete MultiXact scenario and explain how it causes production trouble.
Open the question → - seniorCloud Postgres platformPro answer
When we find a bug we send the patch upstream to pgsql-hackers. Have you worked that way, finding something nobody had flagged and driving it to a fix?
Someone asked whether they have contributed a fix upstream to PostgreSQL, who needs a truthful answer that still demonstrates the skill.
Open the question → - seniorData platformPro answer
n_live_tup and n_dead_tup are estimates. If I need the real number of tuples in a table, what do you reach for?
Someone asked how to get a true tuple count in PostgreSQL and needs to distinguish live rows from physical tuple inspection.
Open the question → - staffData platformPro answer
Explain transaction ID wraparound the way the committer sees it, how does Postgres keep a row from four billion transactions ago still visible?
Someone who needs to explain the wraparound mechanism precisely, including what freezing does and what actually happens at the limit.
Open the question → - midData platformFree in full
You want to measure a table's access pattern cleanly after a change. How do you reset just that table's statistics without wiping the cluster?
Someone who needs to clear one table's activity counters to measure a change, and to explain why that is not the same as ANALYZE.
Open the question → - seniorData platformPro answer
When you propose a new observability signal, how do you justify it, why that metric, and what does it cost to collect?
Someone asked to justify adding a new database metric and needs a decision-first argument rather than a dashboard wish list.
Open the question → - staffData platformPro answer
Start with the highest-severity issue you've personally owned, the one that pushed your limits.
Someone preparing to tell their own worst incident story in a senior or staff interview, who needs the shape rather than a script.
Open the question → - seniorDatabase SRE / Customer ReliabilityPro answer
A customer says “the database is down.” In five minutes, how do you separate a broken connection path, connection saturation, wrong-node routing, lock waits, and a query-specific failure—and what do you communicate before you know root cause?
Someone asked how they would triage a reported database outage in the first five minutes, including what they would say to the customer.
Open the question → - seniorPostgreSQL Performance / MVCC OperationsPro answer
A query became slow after a deploy. Show a safe production triage that separates estimate error, execution or I/O cost, lock waits, and a cleanup horizon. Then explain what VACUUM can reclaim, how freezing protects availability, and how you would prove physical bloat.
Someone asked to triage a post-deploy query regression safely on production, without cancelling the wrong thing.
Open the question → - staffPostgreSQL HA / PatroniPro answer
Patroni leadership loss during a DCS partition: prove one writer.
Someone asked how they would prove a Patroni cluster still has exactly one writer during a DCS partition.
Open the question → - staffBackup / Disaster RecoveryPro answer
Backups are green. Prove an isolated restore to 16:40 UTC yesterday, and show the evidence that establishes both RPO and RTO.
Someone asked to prove a point-in-time restore works, and to say what evidence establishes recovery point and recovery time.
Open the question → - seniorOnline DDL / Index MaintenancePro answer
On PostgreSQL 18, add a defaulted column and create or rebuild an index on a busy 2 TB table. Name the table and index locks, distinguish scans from rewrites, choose the online path, and define abort and cleanup conditions.
Someone asked which locks an ALTER TABLE and an index build take on a large busy table, and how to run them safely.
Open the question → - seniorSecurity / Multi-tenant SaaSPro answer
A fintech tenant must never read or write another tenant’s rows. Show the effective-role threat model across GRANT, RLS, table ownership, referential integrity, and SECURITY DEFINER functions.
Someone asked how they would guarantee one tenant cannot read another tenant's rows, beyond enabling row-level security.
Open the question → - staffPostgreSQL Replication / WAL OperationsPro answer
A PostgreSQL primary's pg_wal is growing while one physical standby and one logical subscriber are behind. Show how you prove which replication slot is retaining WAL, distinguish a slow or stopped consumer from an orphan, decide whether the consumer can catch up, and recover safely without deleting WAL files.
Someone asked how they would diagnose a primary whose WAL directory is growing because of replication slots.
Open the question → - staffPostgreSQL Major UpgradePro answer
You must cut a write workload from PostgreSQL 14 to 18 with logical replication. Where is the rollback boundary, and how do you prove the subscriber caught the source stop LSN?
Someone asked where the point of no return is in a logical-replication major version cutover, and how to prove catch-up.
Open the question →