CockroachDB (pay-as-you-go)

We have been using CockroachDB  for the last two years. From 2023, we started working on a product similar to Google Capture . We decided to work with CockroachDB  with Hasura . Hasura  is a UI for the database CockroachDB that provides a playground to work with the API and data. Since CockroachDB primarily supports SQL, we didn't want to use plain SQL in our programming, so we decided to go with ORM and selected Hasura. With Hasura, we sometimes require GraphQL. Besides using GraphQL, we have also used SQL with CockroachDB directly to get data and perform insert and update operations.

Regarding our operations, CockroachDB equals PostgreSQL  plus Google Spanner , and we can remove scalability, then add resilience across regions.

For CockroachDB multi-active availability, all replicas of our data are active and can serve read and write simultaneously. Currently, we are not using these features because our database is not too large with only 20 customers using our application. MAA is a feature provided by CockroachDB, but they charge for this functionality. It has consequences by raft where each range of data is replicated with default three replicas. Writes go through raft consensus where a majority of replicas must agree, ensuring strong consistency even if some nodes or regions fail.

Regarding multi-region or global clusters, it allows spanning clusters across multiple regions. We are configured with AWS  because we use Lambda functions on the server end. The zone configuration lets you control where replicas live, such as keeping EU customer data in EU and US customer data in US, though we aren't currently using this feature.

ACID transactions are not required in our application because we don't have any payment gateway currently. CockroachDB uses a two-phase commit protocol on top of raft consensus. If any part of the transaction fails, commits, times out, or node crashes, CockroachDB automatically rolls back.

The benefits include being developer-friendly, strong security with RBAC, encryption, and data residency laws by pinning data to specific regions. It is self-healing because if a node crashes, data automatically rebalances to healthy nodes without manual failover, replication, or clustering needed. Finally, it offers cost efficiency.

CockroachDB is excellent for managing relationships in RDBMS . There are multiple benefits such as the ability to create views and dynamic queries that save data and reduce the use of complex algorithms in code. Instead of writing algorithms to get desired output, we can directly call the views where predefined algorithms form the view to get the desired output.

CockroachDB's geo-distribution feature is superior to traditional databases. The admin UI can be used according to requirements. Compared to traditional databases PostgreSQL  and MySQL , CockroachDB provides both SQL features and is horizontally scalable, meaning it is cloud-native. We can start with a single node and scale to hundreds of nodes across multiple regions. It automatically distributes data and workload without manual sharding.

The database offers high availability and fault tolerance with data replicated across nodes and regions. It is compatible with various drivers and ORMs including Prisma, Sequelize, and Hibernate. It provides automatic migrations and online schema changes. Security features include encryption in transit, role-based access control, and compliance with data residency laws by pinning data to specific regions.

CockroachDB supports ACID transactions and is cloud-native, allowing creation of hundreds of microservices with multiple regions. The automatic data and workload distribution eliminates the need for sharding, which is required in databases MongoDB.

The platform should provide a UI feature where users have the ability to manage it. The CockroachDB platform should have a chatbot where users can inquire about pricing or seek solutions for queries, rather than relying on email support.

The current UI only provides SQL playground, which is technical. There should be a UI playground where users with limited technical knowledge can run and update queries, similar to phpMyAdmin or Adminer.

Latency optimization could be improved for globally distributed applications, particularly in query routing. For multi-region deployment, CockroachDB requires at least three plus replicas across data centers to achieve strong consistency across regions, which increases infrastructure costs including compute, storage, and networking.

The disadvantages include cost-efficiency with multi-region deployment and schema migration speed, which should be faster compared to other drivers. The query optimizer should handle complex queries better through the UI interface.

We have been using CockroachDB for the last two years.

I contacted technical support once regarding an error modifying field data types. I received a response within two to three hours, and the issue was resolved efficiently.

Positive

The setup was very straightforward. Everything was mentioned in the documentation, and even a layman or business analyst can complete the setup without having technical knowledge.

The tech support deserves a rating of nine out of ten. My overall rating for CockroachDB is 8 out of 10.

Public Cloud

Amazon Web Services  (AWS)

What do you like best about the product?

We integrated CockroachDB into our existing analytics service stack, which relies on PostgreSQL connections, without encountering any issues. Everything worked immediately without needing any code modifications, workarounds, or special configurations.

What do you dislike about the product?

Based on our experience to date using it for our analytics platform, we haven't identified any dislikes

What problems is the product solving and how is that benefiting you?

We were able to launch our analytics service into production quickly and without initial cost using CockroachDB, which was a major advantage.

What do you like best about the product?

The ease of use offered by the cloud options and the capabilities of the self hosted option.

What do you dislike about the product?

Maybe I lack experience with concurrent use of the DB, but sometimes the default config causes me trouble and I need to change it.

What problems is the product solving and how is that benefiting you?

Currently we have a few products and cockroachdb is the default solution for db we use.

What do you like best about the product?

CockroachDB does what a database should do. Just sit there and hold my data reliably without me having to babysit. No worrying about partitioning. I don't care about servers going up or down or load balancers. I don't care if a new version is available. I just keep my schema up to date. Perform my CRUD operations and get on with building my business.

And did I mention the support? Recently I had an issue where backups to an S3 compatible bucket stopped working. Within 24 hours I was on a call with a product manager talking through the issue to understand exactly what happened and what my recommended path forward would be. Yes, by running a serverless database in the cloud as a non-enterprise I am giving up some level of control, which is what happened here when they upgraded to the latest AWS S3 SDK that had some incompatible headers, but the tradeoff is that I have zero operational headaches to deal with.

The last feature that works perfectly with my business model is the pay-as-you-go model where I can spin up as many clusters and databases as I like without getting nailed for every instance. I am in early days and I'm constantly spinning up new clusters for testing or very small customers. When I can create an isolated cluster for a customer and it only costs me a few dollars a month that's a huge win.

What do you dislike about the product?

The mot difficult thing about working with CockroachDB is learning how to think differently about the operational aspects. You have to accept that you are going to lose some level of control if you want to the Serverless cost effective model. Another challenge I've faced is that although there is a high level of Postgres protocol compatibility there are enough differences that you can't assume a zero effort lift and shift.

What problems is the product solving and how is that benefiting you?

It provides a reliable database with no operational responsibilities for me.

I am studying how to deploy CockroachDB  and YugaByteDB, and learning some basic information about them. I am testing these databases as part of my school application to find a suitable database for our applications. Currently, I am using PostgreSQL , however, I want to try some distributed databases for testing purposes.

The most valuable feature is that CockroachDB  is a distributed database, which can deploy nodes running in different regions. Geo-replication capabilities are also important for production applications. Having database servers close to my customers, for example, in the US or UK, would improve performance.

I would like CockroachDB to have more compatibility with PostgreSQL , especially with the connection string and technical integrations.

I have been studying and testing CockroachDB and YugaByteDB for about six months.

I did not contact the technical support by CockroachDB. However, their website has an AI chatbot, which I found to be very useful.

Positive

Currently, I am using PostgreSQL. I want to try some distributed databases for testing purposes.

I deployed a three-node database in a local system using Docker . I have successfully deployed a secure cluster.

I am not very familiar with the pricing as I am now only in the experimental phase.

I am also studying and testing YugaByteDB alongside CockroachDB.

I'd rate the solution eight out of ten.