Cockroach Labs Blog

Netflix and other innovators are headlining RoachFest 2024, the event series from distributed SQL database CockroachDB. Attend in Menlo Park, London and NYC.

What are the signs that PostgreSQL may no longer be suited for your banking and payments applications and business needs?

Learn how to modernize your database and migrate from Oracle to CockroachDB with this step-by-step guide.

Inside the architecture Starburst built to support exabytes of data across a five-region deployment with CockroachDB that guarantees high availability and low latency access for its customers.

Engineers just want to enjoy their weekend. But outages, hardware failures, and usage spikes can all lead to the dreaded PagerDuty alert, especially when the system being affected is something as critical as your metadata database.

CockroachDB joins the Google Distributed Cloud, an air gapped solution built to serve public sector institutions and highly-regulated enterprises in industries like banking, healthcare, and manufacturing.

This blog explores how to leverage CockroachDB as an OLTP database and BigQuery for OLAP, and then further enhance data capabilities using BigQueryML and Vertex AI for machine learning.

Under READ COMMITTED isolation, PostgreSQL can sometimes miss rows when performing UPDATE, DELETE, SELECT FOR UPDATE, or SELECT FOR SHARE statements. This is due to the EvalPlanQual recheck PostgreSQL adds to these statements to prevent lost-update anomalies.

The Nightfall team runs two major use cases in production on CockroachDB: a metadata storage layer for their API gateway and an interactive analytics dashboard. Read on to learn more.

User data. Application data. Customer history. User account data. Behavioral data. Metadata is at the very heart of your business, and when you do metadata right, people won’t notice you’ve done anything at all.

With the help of Antithesis, a startup from the founders of FoundationDB, CockroachDB engineers were able to track down and reproduce a most elusive one-in-a-million bug, nearly deterministically!

How do you keep databases available and performant at million-QPS scale? Here's how DoorDash does it with CockroachDB, including three tips for success at scale.

You find a mystery server under a desk. It has been sitting there, unpatched, unattended, but running — for the past twenty years. Nobody has the slightest idea of why it's there or what it's doing.

What are the benefits of running a database across multiple clouds, and more importantly: how can we actually accomplish it?

Learn how this sports betting company built a performant, consistent, and compliant mobile app that scales easily to new regions — and saved millions in projected overhead.

Every relational database needs to establish a balance between accuracy and performance. Transaction isolation levels have a direct, though often unrecognized, impact on both.

Learn the difference between an inner join, a left join, a right join, and more. Plus, learn about some special join types in SQL, like self join and cross join.

What are the advantages and disadvantages of managed database-as-a-service offerings, and why should you consider letting someone else operate your database?

Forget the anomaly table. Here's what the different transaction isolation levels mean for the design and implementation of *your* app.

Applications built on a traditional failover pattern with two data centers (and no access to a third on-prem or cloud region) used to face challenges in ensuring high resilience with CRDB. New in v23.2, physical cluster replication functionality now allows limited distribution applications to survive full cluster failures — with RPO and RTO measured in seconds rather than minutes.

How do we know what our database is actually capable of? We test it. A lot. Here's a summary of some of the tests and benchmarks we run regularly to ensure CockroachDB is performant even at global scale.

Inventory management is challenging. Here's a look at how to build modern inventory architecture based on real-world systems built by massive ecommerce companies.

Database migrations are famously difficult. They only happen when the current database is seriously holding back a business, typically in one of three ways. If this is you, CockroachDB's MOLT suite of migration tools is designed to make your database transition as boring as possible.

Cloud storage with enterprise-grade availability, durability, and performance – all for less than what the major CSPs charge? It's possible. Here's how.

CockroachDB v23.2 now seamlessly integrates with Confluent, using change data capture so users can connect the CRDB's distributed SQL capabilities with the real-time event streaming features of Confluent.

CockroachDB 23.2 includes stored procedures/UDFs, performance boosts with smarter data distribution, physical replication, expedited recovery from cloud failures, and Read Committed isolation.

To get you off to a strong start in 2024, here are our top ten most valuable “a-ha” moments and technical takeaways from the past year of Cockroach Labs' Big Ideas in App Architecture podcast.

What is cloud SQL? For that matter, what is Cloud SQL, and is it different from cloud SQL? Let's get to the bottom of this important but somewhat confusing (thanks, Google!) term.

Let's take a look at the most commonly-used data types in Postgres and how they're used in practice, including examples.

It's amazing how a series of small but unfortunate events can quickly add up to one very expensive problem. Like the time one person fat-fingered the wrong year on a mainframe JCL job, Ronald Reagan died, and the US postal service lost $500 million.

How do you build an application that’s scalable, correct, and fast all at the same time? The folks at Materialize know.

Let's take a look at RPO (Recovery Point Objective), RTO (Recovery Time Objective), and learn how to build fault-tolerant applications in which data is never lost.

At RoachFest '23, Shopmonkey Chief Technical Officer Jeff Haynie spoke about the compelling reasons the company moved their core business from Mongo to CockroachDB, and the architectural innovations they implemented to take advantage of CockroachDB’s unique multi-region capabilities.

Each year, the Cockroach Labs blog offers a smorgasbord of posts on technical, educational, informational, and cultural topics to our fellow distributed database devotees (and everyone else, too!). In case you missed them. here are some reader favorites from 2023.

CRDB is shorthand for CockroachDB: the scalable, consistently-replicated, distributed SQL database. A single instance of CRDB can scale from a one laptop to thousands of servers distributed all around the world.

The question comes at you in the hallways, the cafeteria, even the restroom: "Hey, when are WE getting AI???"? It's all anyone wants to talk to you about these days. Here is a handy four-step plan for evaluating and adopting AI so you can satisfy your curious co-workers, instead of simply screaming (on the inside, hopefully).

The crew of the Enterprise was teaching us some important lessons about databases in modern app development, although they didn't realize it.

Modernizing the data layer in your stack can be hugely beneficial, but it can also be risky. Here are three lessons Spreedly learned in their modernization efforts.

Tweeq's cloud-agnostic "no-brainer" tech stack — Go, CockroachDB, Temporal, Bazel, and gRPC, deployed on Oracle Cloud — let them build a payments platform that is scalable to millions of users, available 24/7, extensible, and easy to change.

At RoachFest23, platform engineers Rogger Fabbri and Mario Morgado documented Form3's migration from PostgreSQL and demonstrated the platform's multi-cloud architecture, using CRDB and Kubernetes to span AWS, GCP and Azure with active-active replication to survive any event — even a full cloud outage.

There’s a lot of hype around multi-cloud, but that doesn’t mean adopting a multi-cloud architecture is right for everyone. So what are the "right" reasons to adopt it?

Traditionally, updating your SQL database schema once the database is in production has been a pain. But it doesn't have to be. There's a better way!

True story: It's October, 2012. A company decommissions its DR facility. Days later, Superstorm Sandy hits. You won't believe what happens next!

Running a legacy RDBMS on-prem but dreaming of multi-cloud distributed SQL? It can become reality – but only if you approach the project in the right way.

Aurora Postgres makes scaling reads easy, but writes are limited to a single node, and that limitation took DoorDash's entire application offline for hours.

Learn why Netflix adopted CockroachDB as a scalable SQL database and how they’re using CockroachDB today.

Architects from Hard Rock Digital shares their journey building a sportsbetting and iCasino that can quickly enter new markets with CockroachDB and AWS Outposts.

Black Friday can be a dream for CEOs, but a nightmare for CTOs. Here’s how to build a backend inventory system that’s up to the challenge.

At RoachFest23, Fortinet Director of Software Development Louis Jia shared the story of FortiSASE’s migration from CockroachDB SH (self-hosted) to CockroachDB managed services.

Multi-cloud is not easy. Dig into our free report on the current state of multi-cloud, which contains expert advice from on-the-ground experts on pitfalls to avoid, best practices, and more.

We just released our Browser SQL Shell, which allows writing queries on CockroachDB within seconds of cluster creation.

Schema change happens — a lot. Most relational database systems require taking the DB offline to update your schema. CockroachDB, however, allows online schema changes with no disruption to your application. Here's how it works.

Here's how City Storage Systems built a highly available, highly performant persistence layer that works for mission-critical workloads at global scale WITHOUT needing a massive team.

CockroachDB nodes can be distributed across multiple locations across the globe to ensure that the data can survive outages. Learn about node locality, why this is important in CockroachDB, and how to set this in Kubernetes environments.

Here's how companies using Azure can allow access to CockroachDB for all approved users by dynamically mapping users in Azure Active Directory to a matching SQL user created in CockroachDB.

A serverless database is any database that embodies the core principles of the serverless computing paradigm. This blog details what that means for developers.

The root cause of many outages is human error. At RoachFest23, Thomas Boltze (Head of Cloud and Engineering Excellence with Santander) shared the core tenets of building operationally resilient systems and how to practice them in the real world — fallible humans and all.

The Halloween Problem is a rare but scary issue that can haunt relational databases, mutating your data in unexpected ways if you're not careful.

Learn how to optimize your relational database for better performance, including tuning queries, adding indexes, adjusting your schema, and more!

In this article, we share some key considerations for a technology leader involved in an M&A process and discuss the benefits of choosing flexible, cloud-agnostic technologies.

RoachFest '23, the annual user conference for CockroachDB customers, featured thirty speakers from some of the world’s leading enterprise companies describing the problems they solve, and the innovation they unlock, by using CockroachDB

Learn what kinds of problems often cause high CPU issues in Postgres, and how you can resolve them to improve database and application performance.

When critical tier 0 applications experience disruption or downtime, the outage can result in more than financial losses. Longterm damage to a company's reputation is also a significant risk. Multi-region application architecture mitigates these risks by ensuring high availability.

The cloud is, by definition, a distributed environment. But many traditional databases were created decades ago, in the age of monoliths, and now struggle to provide high availability, horizontal scalability, and performance under modern distributed workloads. So what exactly is a distributed database, and how does it work?