# Lakebase Published articles for Lakebase. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Lakebase Postgres is built on object storage, but 99.99999% of reads never touch it DevFeed: [Lakebase Postgres is built on object storage, but 99.99999% of reads never touch it]() Original publisher: [Read original article]() Author: Carlota Soto Published: 2026-11-05T12:00:00Z Content type: article Language: en Sources: [Neon Docs]() Topics: [Amazon S3](), [Cache](), [data lake]() Tags: [2](), [3](), [about](), [accessible](), [agents](), [all](), [also](), [analytics](), [architecture](), [argument](), [but](), [conclusion](), [engineering](), [isn-t](), [lakebase](), [object](), [object-storage](), [postgres](), [right](), [running](), [storage]() ### AI overview Lakebase Postgres keeps object storage off the active query path through layered caching: the compute node serves about 99.46% of page accesses from its local cache, pageserver NVMe serves about 0.54%, and object storage handles roughly 39 requests per second--about one in 28 million page accesses. Each tier caches data fetched from the tier below. Pageservers answer GetPage@LSN requests by combining WAL deltas with image layers; after base pages were created more eagerly, p99 replay fell to around six deltas, while p90 stayed below one. Object storage fetches mainly occur on first reads after long inactivity or when accessing old branch or restore points; subsequent reads use caches. The architecture also separates compute from durable storage: safekeepers replicate WAL, pageservers persist history asynchronously, and object storage holds immutable history, enabling scale-to-zero, branching, and restores without copying data. ### Source excerpt The most common argument against running Postgres on object storage is that a Postgres backed by object storage must be a slow Postgres. The premise is right, but the conclusion isn't. ## Lakebase Postgres is now generally available on Google Cloud DevFeed: [Lakebase Postgres is now generally available on Google Cloud]() Original publisher: [Read original article]() Author: Cassie Murray Published: 2026-11-02T14:26:08Z Content type: release Language: en Sources: [Databricks]() Topics: [Google Cloud Platform (GCP)](), [Serverless](), [Cloud]() Tags: [access-controls](), [addition](), [agents](), [ai](), [ai-agents](), [ai-platform](), [allows](), [also](), [analytics](), [announce](), [announcements](), [application](), [apps](), [architecture](), [build](), [build-faster](), [cloud](), [google](), [google-cloud](), [lakebase](), [platform](), [postgres]() ### AI overview Lakebase Postgres is generally available on Google Cloud through the Databricks Data + AI Platform. Its architecture separates compute from storage and shares storage with analytical data; the source says this avoids maintaining ETL pipelines and duplicated storage, and brings operational workloads under the platform's shared security and governance model. Databricks presents it as a serverless Postgres foundation for production applications and AI agents, with reduced infrastructure overhead and costs. KaBuM! says it uses Lakebase as an AI-agent memory layer for low-latency reads and writes. To get started, create a Lakebase project from a Databricks workspace on Google Cloud and connect using existing Postgres tools. The excerpt mentions regional availability but does not list the regions or provide detailed reliability and security specifications. ### Source excerpt We're excited to announce that Lakebase Postgres is generally available on Google Cloud... ## Lakebase branching for CI/CD: lessons from 100 PR databases DevFeed: [Lakebase branching for CI/CD: lessons from 100 PR databases]() Original publisher: [Read original article]() Author: Yasamin Tari; Kartik Mathur Published: 2026-10-31T21:25:00Z Content type: article Language: en Sources: [Databricks]() Topics: [CI/CD](), [Pull Request](), [AWS Database Migration Service](), [Replication](), [staging](), [SQL]() Tags: [agent](), [agents](), [ai](), [ai-agents](), [amazon-aurora](), [api](), [audit-logs](), [aws](), [branch](), [cd](), [ci](), [cluster](), [database](), [databases](), [databricks](), [ephemeral](), [industries](), [lakebase](), [pr](), [production](), [technology](), [testing]() ### AI overview Snorkel AI uses Lakebase branches to give each pull request an isolated, writable copy of production data for CI tests. AWS DMS replicates Aurora into Lakebase's main branch, keeping it within about two minutes of production; a PR branch can be created in seconds and is ready in under a minute. Tests run over standard Postgres and can write data or change schemas without affecting production or other PRs. Copy-on-write avoids a full database restore, with storage growing mainly from branch changes (about 5-10 GB per PR in this setup); compute scales to zero when idle. The pipeline deletes branches when PRs finish, with TTL expiration as a cleanup backstop. Production-to-Lakebase and Lakebase-to-CI traffic uses private networking. Developers connect interactively, while CI uses noninteractive service identities; the excerpt also describes short-lived OAuth tokens for agents. Long-lived branches retain older data pages as the main branch updates, so TTLs help limit storage growth. ### Source excerpt The problem: testing against production data shouldn't take over 30 minutesSnorkel... ## Neon Docs search is now built on Neon DevFeed: [Neon Docs search is now built on Neon]() Original publisher: [Read original article]() Author: Andre Landgraf Published: 2026-10-26T12:00:00Z Content type: article Language: en Sources: [Neon Docs]() Topics: [hybrid-search](), [search-engines](), [Embeddings](), [AI-driven information retrieval](), [Test-driven development](), [Algolia](), [CI/CD]() Tags: [about](), [agent](), [agent-experience](), [agent-skill](), [agent-skills](), [ai](), [api](), [blog](), [changelog](), [docs](), [function](), [guides](), [just](), [lakebase](), [neon](), [postgres](), [product](), [reference](), [results](), [search](), [service]() ### AI overview Neon replaced its hosted docs search with a standalone service built on Lakebase Postgres, Neon Functions, and Neon AI Gateway. It stores page chunks, embeddings, and text indexes together; a GitHub Action sends changed pages to a Function, which chunks markdown by heading, embeds it, and updates the database. Search combines BM25 keyword and vector results with reciprocal rank fusion, then reranks them. Because that full pipeline takes seconds, common queries use precomputed, cached hybrid rankings while uncached queries use BM25 alone. The excerpt reports cache latency of 1.9 ms p50 and 2.8 ms p95, and BM25 miss latency of 8.6 ms and 11.5 ms; long-tail queries therefore lack semantic ranking. On 171 real queries, the new system returned the expected page in the top five 94% of the time versus 80% for the previous search, with roughly similar round-trip time. The team also uses a labeled NFCorpus benchmark and post-deploy speed budgets to check regressions. ### Source excerpt We just released a new search for neon.com. Results for the docs, API reference, guides, changelog and blog now come from our own search service: Lakebase Search in Lakebase Postgres, served by a Neon Function. ## How Hamlet uses Databricks Lakebase Postgres to test changes against production data DevFeed: [How Hamlet uses Databricks Lakebase Postgres to test changes against production data]() Original publisher: [Read original article]() Author: Sara Steffen Published: 2026-10-20T14:45:00Z Content type: news Language: en Sources: [Databricks]() Topics: [databricks](), [Pull Request](), [sql-database](), [Artificial Intelligence](), [CI/CD](), [Continuous integration](), [Code review](), [Data Quality](), [AI-Ready Data](), [Data analysis]() Tags: [about](), [accessible](), [actions](), [agent](), [agentic](), [agents](), [ai](), [ai-powered](), [changes](), [company](), [customers](), [data](), [databricks](), [decision-making](), [government](), [lakebase](), [local](), [platform](), [postgres](), [production](), [test](), [uses]() ### AI overview Hamlet creates a Lakebase Postgres branch of production data for each pull request, linked to its preview environment. Engineers and AI coding agents can run pipelines and test fixes against the data that exposed a bug without affecting production; a person still approves every merge, and schema changes reach the parent through code migrations. Hamlet says this replaced 15-30 minutes of staging setup per pull request and several hours per engineer each week spent maintaining test data. Lakebase branches start with the parent's schema and data and use copy-on-write, while standard Postgres database creation copies data. Branch compute scales to zero when idle. Expiring branches are billed for changed data; permanent branches without expiration are billed for their full data size. ### Source excerpt Hamlet, an AI-powered platform that makes local government decision-making accessible... ## Stripe Data Pipeline now syncs real-time Stripe data to Lakebase Postgres DevFeed: [Stripe Data Pipeline now syncs real-time Stripe data to Lakebase Postgres]() Original publisher: [Read original article]() Author: Carlota Soto Published: 2026-10-15T21:08:14Z Content type: release Language: en Sources: [Databricks]() Topics: [real-time]() Tags: [data-pipeline](), [lakebase](), [partners](), [platform](), [postgres](), [stripe]() ### AI overview Stripe Data Pipeline now supports real-time synchronization to Databricks Lakebase Postgres in public preview. Stripe manages backfills, ongoing delivery, and data audits, with updates arriving within seconds. Applications can query the synced data for subscription access, billing details, payment dashboards, and contextual offers. ### Source excerpt If your business runs on Stripe, your application needs access to Stripe data at... ## Sizing Lakebase Search: lessons from a real app and 4 million passages DevFeed: [Sizing Lakebase Search: lessons from a real app and 4 million passages]() Original publisher: [Read original article]() Author: Philipp Tiefenbacher Published: 2026-10-14T21:40:58.161389Z Content type: article Language: en Sources: [Databricks]() Topics: [Qdrant](), [internal developer platform automation](), [Embeddings]() Tags: [ai](), [approximate-nearest-neighbor](), [bm25](), [database](), [lakebase]() ### AI overview The article benchmarks Lakebase Search in a real Databricks app and with public text corpora up to 4 million passages. It describes how compute size, cache ratio, candidate limits, and query settings affect latency, throughput, recall, and cost, and offers sizing guidance based on those measurements. ### Source excerpt Slide Hub is an internal Databricks app. Field engineers use it to find and reuse... ## Lakebase and Agentic SDLC: Branching Databases for Coding Agents DevFeed: [Lakebase and Agentic SDLC: Branching Databases for Coding Agents]() Original publisher: [Read original article]() Author: Thibaut Gourdel Published: 2026-10-08T22:00:00Z Content type: tutorial Language: en Sources: [Databricks]() Topics: [Application Development](), [DeepSeek Harness](), [Continuous integration](), [CI/CD]() Tags: [cloudflare](), [coding-agents](), [continuous-integration](), [databases](), [lakebase](), [platform](), [product](), [sdlc]() ### AI overview The article demonstrates a Lakebase workflow that gives each coding agent and pull request an isolated Postgres database branch for schema changes and tests. It also describes applying migrations through CI, testing preview deployments, and using temporary branches to debug production-derived data safely. ### Source excerpt AI has changed how software gets built. As coding agents take on a growing share... ## Load terabytes of data in minutes into Lakebase Postgres DevFeed: [Load terabytes of data in minutes into Lakebase Postgres]() Original publisher: [Read original article]() Author: Yecheng Yang; Nolan Biscaro; Szu-Po Wang; Pranav Aurora Published: 2026-10-08T18:49:31Z Content type: article Language: en Sources: [Databricks]() Topics: [Supabase](), [Load Balancing](), [table-partitioning]() Tags: [autoscaling](), [b-tree](), [benchmark](), [database](), [databases](), [distributed](), [lakebase](), [postgres]() ### AI overview The article describes Databricks Lakebase Postgres's LTAP architecture for moving large bulk loads from the primary database to distributed Spark processing. It details how Spark-generated heap and index data are validated and staged in object storage before the destination database publishes it, and reports an internal benchmark of loading 1 TB in under five minutes, with index builds not yet parallelized. ### Source excerpt Operational databases like Postgres are built to reliably execute highly concurrent... ## You can outgrow vanilla Postgres without abandoning Postgres DevFeed: [You can outgrow vanilla Postgres without abandoning Postgres]() Original publisher: [Read original article]() Author: Kimberly Fessel Published: 2026-10-08T16:23:19Z Content type: article Language: en Sources: [The New Stack]() Topics: [sql-database](), [Prisma]() Tags: [ai-engineering](), [backend-development](), [columnar](), [data](), [database](), [database-migration](), [databases](), [databricks](), [developer](), [engineering](), [indexes](), [lakebase](), [live-data](), [optimize](), [postgres](), [sponsor-tiger-data](), [sponsored](), [sponsored-webinar](), [tiger-data](), [webinar]() ### AI overview The document promotes an October 21 webinar about how growing live-data workloads can strain vanilla Postgres and possible ways to respond. It describes query optimization, partitioning, autovacuum tuning, and index maintenance, then presents TimescaleDB and Databricks Lakebase as approaches to scaling while retaining Postgres. ### Source excerpt For many data aficionados, Postgres is the way to store and manage relational data. It's efficient, it's straightforward, and it The post You can outgrow vanilla Postgres without abandoning Postgres appeared first on The New Stack. ## BM25 is only as good as your tokens DevFeed: [BM25 is only as good as your tokens]() Original publisher: [Read original article]() Author: Carlota Soto Published: 2026-10-07T12:00:00Z Content type: tutorial Language: en Sources: [Neon Docs]() Topics: [hybrid-search](), [concatenation](), [Query (disambiguation)](), [cosine similarity search]() Tags: [bm25](), [database](), [lakebase](), [neon](), [product](), [search]() ### AI overview The article explains how tokenization shapes BM25 search results in Lakebase Search. It introduces the `lakebase_tokenizer` extension, which lets developers define synonym and stop-word dictionaries in SQL tables, and shows how matching vocabulary can improve ticket search when users use different terms for the same concept. It also describes how token choices affect BM25 relevance scores. ### Source excerpt Lakebase Search is the search primitive of the Neon backend. One of the best things about it is that it ranks keyword results with BM25. That said, BM25 can only score the terms it's given. ## Neon gives you 100 projects for free, with 1 GB of Postgres storage each DevFeed: [Neon gives you 100 projects for free, with 1 GB of Postgres storage each]() Original publisher: [Read original article]() Author: Brad Van Vugt Published: 2026-10-02T12:00:00Z Content type: release Language: en Sources: [Neon Docs]() Topics: [Supabase](), [self-hosted databases vs managed services](), [CI/CD]() Tags: [branch](), [coding](), [coding-agent](), [coding-agents](), [company](), [database](), [lakebase](), [postgres](), [product]() ### AI overview Neon increased storage on its Free plan from 0.5 GB to 1 GB per project, while keeping the allowance at 100 projects; existing projects receive the new limit automatically. The post also describes how the plan supports coding-agent workflows with separate backend projects for experiments and demos. ### Source excerpt Database storage on the Neon Free plan is now 1 GB per project, up from 0.5 GB. You still get 100 projects, with existing projects getting the new limit automatically. ## Lakebase Postgres branch-based restores for fast recovery at scale DevFeed: [Lakebase Postgres branch-based restores for fast recovery at scale]() Original publisher: [Read original article]() Author: Cassie Murray; Carlota Soto Published: 2026-10-01T17:53:52Z Content type: article Language: en Sources: [Databricks]() Topics: [self-hosted databases vs managed services](), [transaction-log](), [Disaster Recovery]() Tags: [database](), [engineering](), [lakebase](), [postgres](), [recovery](), [scale]() ### AI overview The article contrasts traditional Postgres point-in-time recovery, which can require provisioning compute, restoring snapshots, replaying WAL, and waiting for data to hydrate, with Lakebase Postgres branch-based restores. It describes Lakebase's decoupled compute and storage and immutable history, which allow a restore to create a branch at a selected timestamp without copying data. The article says this metadata operation can make past database states queryable quickly, including at large database sizes, and could support product features such as versioning or undo. ### Source excerpt In managed OLTP, restores have always been painfully slow and they get slower at... ## A practical guide to cost optimization with Lakebase Postgres DevFeed: [A practical guide to cost optimization with Lakebase Postgres]() Original publisher: [Read original article]() Author: Benjamin Nwokeleme; Firas Farah; Jen Darrouzet Published: 2026-09-30T16:53:12Z Content type: tutorial Language: en Sources: [Databricks]() Topics: [Heroku Postgres](), [cloud cost management]() Tags: [application-development](), [autoscaling](), [best-practices](), [cost-optimization](), [guide](), [lakebase](), [postgres](), [practical]() ### AI overview This guide explains how Lakebase's separated storage and compute, database branching, autoscaling, and scale-to-zero behavior can reduce costs. Branches share storage for development and testing environments, while adjustable compute limits help align capacity with demand and cap spending. Scale to zero can eliminate compute costs during inactivity, with resumption taking a few hundred milliseconds. ### Source excerpt Lakebase is a fully managed Postgres database, built for the operational realities... ## Lakebase Search: State-of-the-art full text and vector search for Postgres DevFeed: [Lakebase Search: State-of-the-art full text and vector search for Postgres]() Original publisher: [Read original article]() Author: Zhou Sun; Jinjing Zhou; Keming Yang; Usamoi Cui; Pranav Aurora; Junyu Chen Published: 2026-09-28T20:04:33Z Content type: release Language: en Sources: [Databricks]() Topics: [state-management](), [Embeddings](), [Qdrant]() Tags: [autoscaling](), [best](), [database](), [engine](), [full-text-search](), [hybrid-search](), [lakebase](), [pgvector](), [postgres](), [scalability](), [serverless](), [serverless-database](), [vector](), [vector-search]() ### AI overview Databricks announces Lakebase Postgres search extensions for approximate vector search and BM25 full-text search, generally available on AWS and Azure. The article reports benchmark results for vector search and describes customer use of hybrid search across more than 100 million rows. ### Source excerpt Traditional OLTP systems weren't built for the search demands of AI agents. They... ## pgEdge's new agent database branches never merge back into the source DevFeed: [pgEdge's new agent database branches never merge back into the source]() Original publisher: [Read original article]() Author: Amanda Caswell Published: 2026-09-28T14:07:38Z Content type: news Language: en Sources: [The New Stack]() Topics: [AI Context Management](), [database-consistency](), [self-hosted databases vs managed services]() Tags: [ai-agents](), [ai-coding-agents](), [bearer-token](), [cloud](), [cloud-platform](), [database](), [database-branching](), [databases](), [git](), [lakebase](), [mcp](), [model-context-protocol-mcp](), [neon]() ### AI overview pgEdge launched Starfleet, a Postgres cloud platform that gives parallel AI coding agents isolated copy-on-write database branches, with separate credentials and MCP endpoints. Branch changes do not merge back into the source. The report also compares Starfleet with Databricks Lakebase and describes deployment options from hosted service to air-gapped on-premises environments. ### Source excerpt AI coding agents can get an application running quickly, but taking that prototype into production is another problem. The database The post pgEdge's new agent database branches never merge back into the source appeared first on The New Stack. ## Inside branch-based restores in Lakebase Postgres DevFeed: [Inside branch-based restores in Lakebase Postgres]() Original publisher: [Read original article]() Author: Carlota Soto Published: 2026-09-28T12:00:00Z Content type: article Language: en Sources: [Neon Docs]() Topics: [self-hosted databases vs managed services](), [Heroku Postgres]() Tags: [databases](), [failover](), [lakebase](), [postgres](), [product](), [production](), [recovery](), [scale]() ### AI overview The article explains how Lakebase Postgres uses branch creation at a chosen point in database history for restores. Traditional managed Postgres point-in-time recovery provisions an instance, restores a snapshot, and replays WAL, which can take hours at large scale. The branch-based approach points to existing history and attaches independent compute, making a past state queryable quickly; it also describes how agent platforms could use checkpoints to support undo workflows. ### Source excerpt Restores are broken in managed OLTP, and they have been broken for a long time. They are slow, and they get slower with scale. The databases that need recovery most, the large production ones, are the ones left waiting the longest. ## Lakebase, TiDB X, and the Database Architecture AI Demands DevFeed: [Lakebase, TiDB X, and the Database Architecture AI Demands]() Original publisher: [Read original article]() Author: Jinpeng Zhang Published: 2026-09-23T15:35:42Z Content type: comparison Language: en Sources: [TiDB]() Topics: [Database](), [Artificial Intelligence](), [Cloud](), [PostgreSQL](), [databricks]() Tags: [ai](), [cloud](), [database](), [database-architecture](), [database-branching](), [distributed](), [distributed-sql](), [engineering](), [lakebase](), [postgres](), [postgresql](), [tidb-cloud](), [tidb-x]() ### AI overview This comparison examines how Databricks Lakebase and TiDB X separate durable storage from compute for database workloads involving AI coding agents, transactions, branching, and analytical access. It also describes TiDB Cloud Lake as a complementary analytical service and notes production questions that remain unresolved. ### Source excerpt An AI coding agent can generate an application, change its schema, test several implementations, and discard most of them within a single session. Another agent might spend that session updating orders and querying months of operational history. Both need durable state, correct transactions, and fast access. They also need something query throughput doesn't measure: databases [...] The post Lakebase, TiDB X, and the Database Architecture AI Demands appeared first on TiDB.