A set of properties that guarantee database transactions are processed reliably: Atomicity (all operations succeed or fail together), Consistency (data integrity is maintained), Isolation (concurrent transactions don't interfere), and Durability (committed changes persist).
Read full glossary pageAgent-ready Postgres is a PostgreSQL operating model designed so AI agents can use database context and tools inside controlled, testable workflows.
Read full glossary pageAI data governance for Postgres is the set of controls for how AI systems access, test, transform, and audit PostgreSQL data.
Read full glossary pageAI database branching is the practice of creating isolated database branches for testing AI features, agent workflows, retrieval changes, and generated SQL.
Read full glossary pageConfiguration parameter that specifies the shell command to execute to archive a completed WAL segment.
Read full glossary pageConfiguration parameter to enable WAL archiving. Values: off | on | always (for standbys).
Read full glossary pageContinuous shipping of WAL segments to durable storage for backups and PITR. Controlled by archive_mode and archive_command.
Read full glossary pageThe property that ensures all operations within a transaction are completed successfully, or none are applied at all.
Read full glossary pageAutovacuum is PostgreSQL's background maintenance daemon that reclaims storage from dead rows, updates table statistics, and prevents transaction ID wraparound — keeping tables healthy without manual intervention.
Read full glossary pageA tree data structure that maintains sorted data and allows searches, sequential access, insertions, and deletions in logarithmic time.
Read full glossary pageLogical backups dump SQL/data (pg_dump), while physical backups copy data files/WAL (pg_basebackup, file system snapshots).
Read full glossary pageA consistent physical copy of a PostgreSQL cluster, typically taken with pg_basebackup, used with archived WAL for PITR.
Read full glossary pageBYOC is a deployment model where a vendor-managed software layer runs in the customer cloud account or infrastructure boundary.
Read full glossary pageCardinality is the number of distinct values in a column or the estimated number of rows a query step will return — the PostgreSQL query planner uses these estimates to choose between scan and join strategies.
Read full glossary pageA point in the WAL where all dirty data pages have been written to disk, providing a consistent recovery point.
Read full glossary pageA database clone is a separate database copy or branch created from an existing source so teams can read and write without changing the source.
Read full glossary pageConnection pooling reuses a smaller set of database connections across many application requests to reduce connection overhead and protect Postgres.
Read full glossary pageCopy-on-write is a technique where data is shared until a write occurs, then only the changed data is copied or stored separately.
Read full glossary pageDatabase branching is a workflow for creating isolated, writable database copies from a known source so teams can test changes without affecting the main database.
Read full glossary pageDatabase platform engineering is the practice of building reusable, governed database workflows for developers, QA teams, and operators.
Read full glossary pageA deadlock occurs when two or more transactions each hold a lock the other needs, causing them to wait indefinitely until PostgreSQL detects and resolves the cycle by aborting one transaction.
Read full glossary pageThe ACID property ensuring that once a transaction is committed, it remains so even in the event of system failure.
Read full glossary pageA PostgreSQL command that shows the execution plan of a query along with actual runtime statistics.
Read full glossary pageExtended statistics are manually declared statistics objects in PostgreSQL that capture correlations between multiple columns, helping the query planner make accurate row estimates for multi-column conditions.
Read full glossary pageA constraint that establishes a link between data in two tables by referencing the primary key of another table.
Read full glossary pageA technique for searching text data that goes beyond simple pattern matching to include linguistic analysis.
Read full glossary pageGeneralized Inverted Index - an index type in PostgreSQL optimized for indexing composite values like arrays and full-text search.
Read full glossary pageA standby server that can accept read-only queries while continuously applying WAL records from the primary.
Read full glossary pageHTAP is an architecture pattern that supports transactional and analytical workloads closer together instead of separating them into fully isolated systems.
Read full glossary pageA data structure that improves the speed of data retrieval operations on a database table.
Read full glossary pageThe ACID property that ensures concurrent transactions don't interfere with each other.
Read full glossary pagePostgreSQL data types for storing JSON documents. JSONB is a binary format that supports indexing and efficient operations.
Read full glossary pageKubernetes-native Postgres is a PostgreSQL operating model that integrates database lifecycle, automation, and platform controls with Kubernetes concepts.
Read full glossary pageA mechanism to control concurrent access to database resources to maintain data consistency.
Read full glossary pageA method of replicating data changes based on their logical representation rather than physical storage.
Read full glossary pageA byte position in the WAL stream that uniquely identifies a point in time for recovery or replication.
Read full glossary pageA database object that contains the result of a query and is physically stored, unlike a regular view.
Read full glossary pageMVCC is PostgreSQL’s concurrency model that lets transactions see consistent row versions while readers and writers operate at the same time.
Read full glossary pageOLAP is a workload pattern optimized for analytical queries, aggregations, reporting, and exploration over larger datasets.
Read full glossary pageOLTP is a workload pattern optimized for many small, correct transactions such as inserts, updates, reads, and commits from applications.
Read full glossary pageAn OpenShift database platform is an internal platform for provisioning, operating, and governing databases on or around Red Hat OpenShift.
Read full glossary pageA technique to split large tables into smaller, more manageable pieces based on specific criteria.
Read full glossary pagePerformance baseline testing establishes known-good query timing benchmarks for a PostgreSQL database, enabling regression detection when schema changes, index modifications, or query rewrites alter execution behavior.
Read full glossary pagepg_basebackup is a PostgreSQL utility that creates a physical base backup of a running database cluster, often for replicas or PITR.
Read full glossary pagepg_dump is a PostgreSQL utility that creates a logical backup of a database as SQL or a portable archive format.
Read full glossary pagepg_restore is a PostgreSQL utility that restores databases or objects from non-plain-text archive files created by pg_dump.
Read full glossary pagePostgreSQL's built-in benchmarking tool for running performance tests with simulated workloads.
Read full glossary pagePhysical backups copy data files/WAL (cluster-level); logical backups export SQL or data at the database/table level.
Read full glossary pagePITR is a PostgreSQL recovery method that restores a database to a specific point in time using a base backup plus archived WAL records.
Read full glossary pageA Postgres data platform is a platform layer around PostgreSQL that standardizes database lifecycle, development, QA, analytics, and operational workflows.
Read full glossary pagePostgres for AI agents is the use of PostgreSQL as a durable operational data layer for AI agents, retrieval workflows, and agent-driven application changes.
Read full glossary pagePostgres on OpenShift means running PostgreSQL workloads and database platform services on Red Hat OpenShift or an OpenShift-based private cloud.
Read full glossary pageA Postgres vector database is a PostgreSQL-based system that stores and queries embeddings alongside relational data, metadata, and SQL workflows.
Read full glossary pageA prepared statement is a server-side query template that PostgreSQL parses and plans once, then executes multiple times with different parameter values — reducing overhead and preventing SQL injection.
Read full glossary pageA constraint that uniquely identifies each row in a table and cannot contain NULL values.
Read full glossary pagePrivate cloud Postgres is a PostgreSQL operating model where databases run inside a customer-controlled private cloud or dedicated infrastructure boundary.
Read full glossary pageA query execution plan is the step-by-step strategy PostgreSQL's planner selects to retrieve or modify data, showing which indexes, join methods, and scan types will be used.
Read full glossary pageThe component of PostgreSQL that determines the most efficient way to execute a given SQL query.
Read full glossary pageRAG with Postgres is a retrieval-augmented generation pattern that uses PostgreSQL to store, filter, and retrieve data used by AI applications.
Read full glossary pageReal-time analytics is the practice of analyzing fresh operational data with minimal delay so teams and applications can act sooner.
Read full glossary pageA recovery stop point defined by time, XID, name, or LSN to which PostgreSQL will recover during PITR.
Read full glossary pageThe process of copying and maintaining database objects in multiple database environments.
Read full glossary pageA mechanism that retains WAL required by a subscriber or backup stream to prevent premature recycling.
Read full glossary pageProcess of rebuilding a database or cluster from backups. Logical restore uses pg_restore; physical restore replays WAL.
Read full glossary pageConfiguration parameter that specifies how the server retrieves archived WAL files during recovery.
Read full glossary pageRow-Level Security is a PostgreSQL feature that restricts which rows a user or role can access in a table based on a policy, enforcing data isolation at the query level.
Read full glossary pageA self-hosted Postgres platform is a customer-operated platform layer for provisioning, governing, and automating PostgreSQL environments.
Read full glossary pageServerless Postgres on-prem is an operating model that brings self-service, elastic database provisioning to on-premises or private-cloud PostgreSQL environments.
Read full glossary pageSnapshot isolation is a concurrency control mechanism where each transaction reads from a consistent point-in-time snapshot of the database, preventing dirty reads and non-repeatable reads without blocking writers.
Read full glossary pageSovereign Postgres is a PostgreSQL operating model designed to keep data, control planes, and operational access aligned with sovereignty or residency requirements.
Read full glossary pageA standardized language for managing and manipulating relational databases.
Read full glossary pageA standby.signal file in the data directory triggers standby mode in v12+, replacing recovery.conf.
Read full glossary pageAn open-source backend platform that provides PostgreSQL database, authentication, and real-time subscriptions.
Read full glossary pageData organized by time order, typically used for metrics, logs, and sensor data.
Read full glossary pageA sequence identifier for WAL history used during recovery and replication; increments on divergence (e.g., PITR).
Read full glossary pageA key performance metric measuring the number of database transactions completed per second.
Read full glossary pageA sequence of database operations that are executed as a single logical unit of work.
Read full glossary pageTransaction isolation levels define how visible the changes made by one transaction are to other concurrent transactions, trading consistency guarantees against performance and concurrency.
Read full glossary pageA unified database approach keeps more application, analytical, search, and AI workflows close to a common database foundation instead of splitting every use case into a separate system.
Read full glossary pageAn operation that inserts a new row or updates an existing row if a conflict occurs.
Read full glossary pageA PostgreSQL maintenance operation that reclaims storage and updates statistics for optimal performance.
Read full glossary pageVector search finds similar items by comparing embedding vectors rather than matching only exact keywords or relational predicates.
Read full glossary pageVela is a Postgres data platform for creating production-like database branches, clones, and developer workflows around familiar PostgreSQL.
Read full glossary pageVMware exit for Postgres is the process of moving PostgreSQL operations and database workflows away from VMware-centered infrastructure to a newer platform model.
Read full glossary pageWAL is PostgreSQL’s write-ahead log, a durability mechanism that records changes before they are applied to database data files.
Read full glossary pageA SQL clause used to filter records based on specified conditions.
Read full glossary page