Couchbase Sync Gateway

We are using RAG here, and Couchbase Server  is supported as a vector store, providing a high-performance solution for managing and retrieving vector data. This integration helps us to create a no-code AI workflow with flexible vector storage. Flowise is an open-source tool that allows us to create and manage workflows with a visual no-code interface. We install Flowise using NPM  and then use Docker  to deploy our project for running it in a container. We upsert into Couchbase Server  and run queries using a Couchbase Server retriever. We process and generate data using ChatGPT 4O and handle multi-retrieval using multi-interaction to ensure accurate and context-specific responses.

We have used Couchbase Server in our applications, including a sales plus application. In that application, we moved card validation from the Xamarin code and created a tool to validate all card validations. We directly interact with the document and check each document in Couchbase Server to ensure it is correct. Sometimes we pull data from Couchbase Server, but some attributes are not present. We created a tool to validate the data, and when orders are placed, payment is processed and then shipped, with Couchbase Server storing the state of the active workflow as a JSON document. In our project, we use the same approach. We run a distributed workflow, and if it fails, we can directly fetch the data from Couchbase Server to see how it is handling the data between our microservices without forcing them to communicate directly.

I have used Couchbase Server for approximately four to five years. We interact with Couchbase Server and extract data every day. We created a tool named JSON Copilot tool, which is used for schema conversion and automated validation. Sometimes documents are not proper when we create a table, so we validate the data using this tool. We can directly take the document from Couchbase Server and validate it in our tool. This is very useful for extracting and querying JSON documents without a rigid schema.

We have used a vector approach to get data using the RAG model. It also supports offline data, which is very useful for our team for high-performance applications, caching, and generative AI.

We use Couchbase Server as an eventing service so that the data service can initiate and advance the workflow based on raw data. When we integrated it into the workflow system, Couchbase Server became the center of the data lifecycle. It helps us validate data, and when client applications update a document, Couchbase Server triggers JavaScript-based logic, such as on update and on delete, and instantly forwards that payload to an external orchestrator or message broker. This is very useful for our case.

Couchbase Server's best feature is that it operates directly with no-code workflow builders for modern AI workflows. We support vector store integration, which means we can use an embedded vector knowledge field that can be used by AI to inject semantic search context into the LLM

We could use hyperscale vector indexing and upgrade the storage engine so that density and throughput would be higher. If Couchbase Server introduced natural language query capabilities, that would be helpful for everyone to use it in a more effective manner. The logging part could also be improved.

Advanced security and logging would help us make the application more resilient.

Couchbase Server is the best, but if natural language query capabilities were introduced, it would be easier for developers. Instead of writing queries, they could write using natural conversion language and give prompts to Couchbase Server, and they could get results in seconds.

We have used this solution for approximately five years with AI development.

Couchbase Server is very stable.

Couchbase Server allows our organization to independently scale specific data services horizontally and vertically without downtime.

We have a support team in our project, and they are very helpful. Whenever we encounter an issue, they immediately work to have a solution ready within 30 minutes based on severity if it is a P1 or P2 issue. We have not faced many issues with the Couchbase Server support team. We open ticket escalations, they review them, and then they fix the issue.

We have used Postgres, but we found Couchbase Server to be more scalable and perform better. We have chosen to use this database.

We have evaluated MongoDB and Postgres server. We found Couchbase Server very useful for our application since we are supporting NoSQL documents and validating data. Performance-wise, writing queries is very easy and comfortable in Couchbase Server across distributed applications.

Couchbase Server supports distributed transactions, which ensures data integrity by supporting distributed transactions across the database. It has a decentralized architecture, which prevents single points of failure and enables seamless scaling and disaster recovery.

If you want to adopt Couchbase Server, we should take full advantage of Couchbase Server query language, which allows developers to use familiar SQL syntax to query unstructured and semi-structured JSON data. Leveraging JSON flexibility by designing documents into nested arrays and flattening data together can minimize multi-document joins. This is very useful for using SQL schemas with tables as collections and complex foreign keys. Couchbase Server also has a memory-first architecture that reads data directly from RAM before committing data to disk. It tunes for the active data set. I would recommend every developer to use it. It uses an autonomous operator and tests rebalancing on load. If you want to adopt Capella, there is a team that will fully manage the Capella database as a service to eliminate operational overhead. I would recommend everyone to use this.

Couchbase Server is a high-performance NoSQL document database. We primarily use it for caching and generative AI to develop modern applications using rapid data access and offline support for offline mobile applications, along with flexible data modeling. We have used it for dynamic product catalogs, user profiles, and web applications.

When we started using Couchbase Server, it achieved very good critical business and operational results. The major outcome is consistent microsecond to sub-millisecond response times from read and write operations during periods of concurrent requests. Couchbase Server uses a memory-first architecture, which is very easy for integrating the caching layer. Data is read from and returned to RAM before being persisted to disk.

Couchbase Server has a security architecture that enforces strict data governance and regulatory compliance across distributed edge topology. There is TTL governance that automatically expires sensitive records natively at the document level.

Couchbase Server version 8.8.0 has achieved a balanced combination of operational throughput and data precision up to 93% accuracy while maintaining performance.

Couchbase Server is deployed as a database as a service in a hybrid cloud environment. I would rate this review an 8.

Couchbase Server 's main use case for our organization is storing data in JSON format where we have the document ID we need to store. If data is in JSON form, we can use Couchbase Server , and it will provide a cache layer before storing the data. We are primarily using Couchbase Server to store JSON data itself, but there are many more usages. It is a NoSQL distributed database built on low latency and high throughput. We can store various data such as user sessions, preferences, and profiles.

Couchbase Server is particularly useful for user sessions and preferences. A session is commonly used to store temporary data for login purposes. Whenever a user enters their credentials, we first validate them. If validation is successful, we generally store the session information in the backend. The data that the user provides is in JSON form, where we can set the user ID, username, login time, and all related information. We store the data using a key prefix, which is session colon session ID, with the session ID generated in the Java backend. We set a timeout, and if we set any timeout, such as 50 minutes, the session will expire after that time, requiring the user to send the data again. We are also using an index to store query information.

In my current project, I have been working in the airline domain, so if a user is changing or rescheduling any ticket information, we need to send all modification information to our timers. Currently, we are using Couchbase Server as the primary database, where we need to store settings to perform operations in our project. We store settings in a way that identifies the records with a document ID provided by Couchbase Server. Couchbase Server's primary benefit is that our data is in JSON form, and it is very fast in comparison to saving and retrieval with low latency. We don't need to write extensive queries to fetch data; we write any Couchbase Server repository, and if we add that find by ID or something similar, it will fetch everything in less time.

Couchbase Server offers several valuable features, including the ability to store flexible JSON documents with any structure, and it provides an in-memory architecture with first reading, then writing capabilities. We can apply auto-scaling in Couchbase Server by adding multiple nodes, which automatically distribute the data across nodes while having built-in cache. When Couchbase Server is built, they provide a disk to store and cache together, so if data is in the cache, it will not hit the database directly; instead, it will fetch data from the cache.

Whenever we use SQL Server , we use queries to fetch data. Similarly, Couchbase Server provides N1QL queries for data retrieval. In terms of security features, Couchbase Server offers role-based access control with automatic encryption and TTL timeout settings for automatic destruction upon reaching the time limit. The security capability includes role-based access control at the bucket level and essentially at the scope level, granting access based on defined roles. It also supports various authentication types, including built-in username and credential systems and LDAP.

Couchbase Server has positively impacted my organization as every company deals with data, which we are storing in Couchbase Server. When a user interacts with Couchbase Server for the first time, it fetches data from the database and stores it in the cache, so subsequent hits fetch directly from the cache without going to Couchbase Server. We don't need to explicitly set any caching technology such as Redis  or Hazelcast since it has its own cache, making Couchbase Server retrieval for data purposes comparable to others.

With one year of experience, I can suggest improvements for Couchbase Server. From what I have used, it is pretty good compared to other databases, especially since our data is in JSON format. It would be a great improvement for Couchbase Server to support various data structures beyond JSON, as supporting heavy complex queries would enhance its competitive edge. Allowing different formats would be beneficial.

Regarding needed improvements, I can note any technology associated with Couchbase Server. Recently, I used Java with Couchbase Server in my project, where Java provides documentation that we can utilize. Currently, we are embedded with Couchbase Server, and Spring  has made a Couchbase Server repository for integration purposes.

To improve user experience, it would be beneficial for Couchbase Server to support various data structures, enabling compatibility with formats such as PDF, JSON, XML, and others.

Supporting every kind of data structure would be a substantial improvement for Couchbase Server.

Recently, I have used Couchbase Server in my current company, and I have served one year completely in Couchbase Server.

In my experience, Couchbase Server is stable. While complex queries may take time, I have not encountered any significant issues, and I can say it is very stable as of now.

Couchbase Server's scalability meets our needs, handling growth and changes in workload well. We employ auto-scaling based on memory utilization and health, enabling it to automatically scale, which is very helpful.

I changed organizations, where Couchbase Server is used because our data is in JSON format and structured accordingly. In my previous role, we utilized a SQL database as our data was interdependent, but now, the JSON format is sufficient for our needs. The choice ultimately depends on the structure of data since relational databases are better suited for data with inherent relationships.

Whenever we are dealing with a dashboard, we can use time annotations to calculate performance metrics. Currently, we are using GCP  as a cloud service. We can add time rotation, so if we want to check performance or method outputs, we can utilize system.current time to measure the start and end time for the query operation, which depends on the complexity of the method. Optimized methods yield results in milliseconds, while more complex queries might take longer. I cannot provide an exact estimate, as it all depends on the data complexity. Using an index for frequently accessed data can be helpful since fetching from cache is faster compared to other methods.

I have not used Couchbase Server's AI capabilities yet. Currently, we are focused on storing and retrieving data faster with Couchbase Server, but I haven't had the chance to explore AI features.

I haven't utilized the AI capabilities of Couchbase Server, so I'm uncertain about its accuracy or reliability in output. However, Couchbase Server can handle billions of records, fetching them within milliseconds, but without firsthand experience in AI, I cannot provide thoughts on that aspect.

My advice for others considering Couchbase Server is to determine if the data is structured in JSON format, as it offers significantly improved facilities compared to others. I would rate this product an eight out of ten.

I use Couchbase Server  as a main database to store information as key and value objects for geolocation-based lookups for place data. I use it for many purposes, depending on the use case. I use its indexing feature, its search feature, its querying feature, and everything depending on the type of project and complexity. Halalfoodle is a location information catalog for restaurants. I also have another use case for an ORTB system to keep analytical counters, not all the analytical data, but counters and configurations, because it is easy to set up and easy to manage with its beautiful UI, which saves me from needing extra UI tools. It saves me time.

I use Couchbase Server because it allows me to use it as flexible data storage. If I want, I can use it as key-value storage. If I want, I can use the SQL feature to query it. If I want, I can use its analytical views feature to create necessary views to query. I think its primary and greatest feature is the UI, which simplifies many things. With another database, you have to have a separate client utility from the terminal. This one gives you a UI which is beautiful and easy to use. It really saves time. The XDCR feature is the best. I simply set it up, define the IP address, and it automatically distributes the data between the data center instances, which is a very helpful feature. There is no headache, everything is out-of-the-box, and all of it is great.

Couchbase Server has definitely improved the workflow in my organization. The UI out-of-the-box does many things. I recently discovered collections and scopes, which I was not aware of, but after discovering them, they saved me many things related to governing indexes and creating special indexes to query. Now I can put data structures in the collections, which I can directly look up. I know that in a specific collection, I will get only specific types of documents, instead of having everything together in one place. That is great. After struggling with custom-based sequences and custom pointer documents that have numeric values pointing to the original document, now with collections and scopes, I can separate data structures, which saves me time. The UI is the best. All of these features have saved me maintenance costs and improved operation performance. I can quickly log into the UI and handle all the necessary operations. I can check logs and see warnings, everything in one place. I do not need extra items to use it.

Couchbase Server is very performant. I had a situation where I used it as session storage once in my lifetime, and it was very powerful. It handled DDoS situations when the application was creating 100,000 sessions in a second, and it was very fast.

It would be great to have a kind of pub/sub in Couchbase Server. I want to have a kind of stream in the UI to create a stream separately and have some metrics. By pushing the data to the stream, I could consume it on another side. Currently, I am using Redis  or Redis-based applications. I use NATS from GoLang. I use Kafka and RabbitMQ. I need something out-of-the-box that is performant as Couchbase Server. It will save me time from having extra items to be installed. With Couchbase Server, everything is centralized, whatever I need, even to create triggers on some kind of collections when something changes, it triggers something, and I can immediately process it and do everything else.

Another thing I would require comes from CouchDB . In CouchDB , I was able to have multiple versions of the same document. It is like a document, and you have revisions of the document. In Couchbase Server, I know it is built upon CouchDB, and I see the metadata and other related elements, but I want to have the easiest way to see the versions of the document from the UI or from the client-side, just querying to show me revisions and just taking them by revision, rolling back, or whatever it is. I know everything is existing in the protocol and under the hood, but making such features easier to use and visible would be great. Pub/sub, or streams and data streams, and the ability to see the revisions of documents would be a great addition.

I have been using Couchbase Server since 2014.

Couchbase Server is stable enough in my experience.

I use XDCR in Couchbase Server to add servers and do the master replication. It works beautifully out-of-the-box.

I have not contacted customer support for Couchbase Server because I use the community version. I am on my own and do my own project. For some other components, I always use community-based versions. That is why I have not used support.

I previously used MongoDB and ArangoDB before choosing Couchbase Server. All of them are good. From knowing how Couchbase Server works and the performance outcome which I saw in real projects during battle testing and during anti-DDoS efforts, during my own project, I still keep using Couchbase Server and I am comfortable with it. I never think about switching to SQL-based servers. I do not think about switching to MongoDB and others. This is only the case if I am able to control the architecture.

It is easy to install with no problem.

That is the thing. I do not need DevOps or DBA database admins. Being a CTO of my own project, I simply installed Couchbase Server in multiple servers, configured XDCR, connected any of the server instances, and operated with them. It did everything related to automatically distributing the data out-of-the-box. I am currently building a client driver for the Crystal language stack, which is helping me to understand how Couchbase Server works under the hood by utilizing Memcached protocol and HTTP-based protocols. I am getting familiar with Couchbase Server, which makes me learn new things. By doing it, I am getting more excited about how it makes a developer's life simpler, and there is no need for an extra person to support it. I simply go, install, configure XDCR, check the logs, check necessary parameters, and everything is out-of-the-box. Nothing specially is needed. I even use the CLI tools for backing up and recovering. It is a time-saver. Everything I need is out-of-the-box. I do not know what else I could ask for. It impressed me. I am more than happy.

It is very opinionated because every other database is competing with each other and adding extra new features to be successful. That is why it is a bit hard to give some concrete opinion. My vision for this is if you are experienced enough and you need something flexible where you can enable and disable some features, by keeping key-value storage only or adding SQL++ feature, or adding views, or adding search features, or indexing, Couchbase Server is giving all flexibilities. You can always correctly define how to use memory and how to distribute computer servers' powers. I would give this product a rating of 8.