The PostgreSQL administration field guideField notes · Runbooks · Free certification

10 incident labs · safe simulated systems

Refresh the judgement behind the credential.

Practice the next observation in Guided refresh or Challenge refresh mode. Make decisions under a simulated clock, then use the evidence and feedback to rehearse the relevant certification competency whenever the skill needs refreshing.

These repeatable incident drills run in the browser and are separate from the assessed real-SQL module labs. Use them whenever you want a quick refresher; they do not change your certification record.

Review the certification competencies →

Rebuild the operating model

Read each explanation as you go and follow the strongest evidence boundary.

Test recall under a clock

Commit to the next observation first, then use feedback as your debrief.

Case 01

The replica that stopped replaying

A four-phase investigation that separates impact, transport, replay, and safe response.

advanced · 12 minOpen case →
Case 02

The migration at the front of the queue

A compact exercise in finding the root of a blocked-write fan-out.

practitioner · 4 minOpen case →
Case 03

The cheap query that became expensive

A compact exercise in ranking statements by aggregate load.

primer · 4 minOpen case →
Case 04

The table vacuum cannot clean

A compact exercise in distinguishing vacuum capacity from a pinned cleanup horizon.

practitioner · 5 minOpen case →
Case 05

The pool that consumed every backend

Find the application and session state consuming connection capacity before moving the limit.

practitioner · 6 minOpen case →
Case 06

The volume with ninety minutes left

Separate relation, WAL, temporary, and log growth before attempting reclamation.

advanced · 6 minOpen case →
Case 07

The backup that stopped completing

Diagnose archive, storage, permission, and duration evidence without declaring protection from stale output.

practitioner · 6 minOpen case →
Case 08

The restore rehearsal that missed its objective

Treat restore time, dependencies, WAL, and validation as parts of one recovery path.

advanced · 6 minOpen case →
Case 09

The checkpoint that saturated storage

Correlate checkpoint write pressure with latency before changing durability or memory settings.

advanced · 6 minOpen case →
Case 10

The safe-looking deploy that changed the plan

Connect deploy timing, parameter shapes, plans, locks, and rollback compatibility.

practitioner · 6 minOpen case →