A newer version of this documentation is available.

View Latest

Sizing Guidelines

      Evaluate the overall performance and capacity goals that you have for Couchbase, and use that information to determine the necessary resources that you’ll need in your deployment.

      When you plan to deploy a Couchbase Server cluster, perhaps the most common (and important) question that comes up is: How many nodes do I need and what size do they need to be?

      With the increasing number of Couchbase services and the flexibility of the Couchbase Data Platform, the answer to this question can be challenging. This guide aims to help you better size your deployment.

      If you want detailed recommendations for your specific deployment, you can contact Couchbase Support.

      The sizing recommendations and calculations discussed in this guide are based on an analysis of performance data and common use-cases.

      General Considerations

      The sizing of your Couchbase Server cluster is critical to its overall stability and performance. While there are some basic system requirements to run Couchbase Server, you still need to evaluate the overall performance and capacity requirements for your workload and dataset, and then divide that into the hardware and resources you have available.

      Your application wants the majority of reads to come out of the cache, and to have the I/O capacity to handle the writes. There needs to be enough capacity in all areas to support everything the system is doing while maintaining the required level of performance.

      Multi-Dimensional Scaling

      Couchbase Services are what allow you to access and maintain your data. These services can be deployed, maintained, and provisioned independently of one another. This independent service model allows you to take advantage of Multi-Dimensional Scaling, whereby a cluster can be fine-tuned for optimal handling of emergent workload-requirements, on a service-by-service basis.

      Since each service has different demands on hardware resources, Multi-Dimensional Scaling plays an important role when sizing your Couchbase cluster, both pre and post-deployment. For example, core Data Service operations can often benefit greatly from the scale out of smaller commodity nodes, whereas low latency operations with the Query Service may see a greater benefit from the scale up of hardware resources on a given node.

      For more information about the nature and resource demands of each Couchbase Service, refer to Services.

      About Couchbase Server Resources

      This guide discusses four types of resources that you should consider when sizing a Couchbase Server cluster node:


      CPU refers to the number of cores and clock speed that are required to run your workload.


      RAM is frequently one of the most crucial areas to size correctly. Cached documents allow the reads to be served at very low latency and consistently high throughput.

      This resource represents the main memory that you allocate to Couchbase Server and should be determined by the following factors:

      • How much free RAM is available beyond OS and other applications

      • How much data do you want to store in main memory

      • How much latency you expect from KV/indexing/query performance

      Some components that require RAM are:

      • Memory-optimized Global Indexes, which enable in-memory index processing and index scans with the lowest latency.

      • Full Text Search (FTS)

      Table 1. Minimum RAM Quota for Couchbase Server Components
      Component Minimum RAM

      Data Service

      1024 MB

      Index Service (Standard Global Secondary)

      256 MB

      Indexing Service (Memory-Optimized)

      256 MB minimum, 1024 MB and above recommended

      Search Service (Full-Text Search)

      256 MB minimum; 2048 MB and above recommended

      Query Service

      No RAM-allocation is required for this service.

      Eventing Service

      256 MB

      Analytics Service

      1024 MB

      Storage (disk space)

      Requirements for your disk subsystem are:

      • Disk size — which refers to the amount of the disk storage space that is needed to hold your entire data set.

      • Disk I/O — which is a combination of your sustained read/write rate, the need for compacting the database files, and anything else that requires disk access.

      To better support Couchbase Server, keep in mind the following:

      • Disk space continues to grow if fragmentation ratio keeps climbing. To mitigate this, add enough buffer in your disk space to store all of the data. Also, keep an eye on the fragmentation ratio in the Couchbase user interfaces and trigger compaction processes when needed.

      • Solid State Drives (SSDs) are desired, but not required. An SSD will give much better performance than a Hard Disk Drive (HDD) when it comes to disk throughput and latency.


      Enough network bandwidth is vital to the performance of Couchbase Server. A reliable high-speed network for intra-cluster and inter-cluster communications has a huge effect on overall performance and scalability of Couchbase Server.

      Most deployments can achieve optimal performance with 1 Gbps interconnects, but some may need 10 Gbps.

      Sizing Data Service Nodes

      Data Service nodes handle data service operations, such as create/read/update/delete (CRUD).

      It’s important to keep use-cases and application workloads in mind since different application workloads have different resource requirements. For example, if your working set needs to be fully in memory, you might need large RAM size. On the other hand, if your application requires only 5% of data in memory, you will need disks with enough space to store all of the data, and that are fast enough for your read/write operations.

      You can start sizing the Data Service nodes by answering the following questions:

      1. Is the application primarily (or even exclusively) using individual document access?

      2. Do you plan to use Views?

      3. Do you plan to use XDCR?

      4. What’s your working set size and what are your data operation throughput and latency requirements?

      Answers to the above questions can help you better understand the capacity requirement of your cluster and provide a better estimation for sizing.

      The following is an example use-case for sizing RAM:

      Table 2. Input Variables for Sizing RAM
      Input Variable Value




      100 bytes


      10,000 bytes





      Table 3. Constants for Sizing RAM
      Constants Value

      Type of Storage





      56 bytes



      Based on the provided data, a rough sizing guideline formula would be:

      Table 4. Guideline Formula for Sizing a Cluster
      Variable Calculation


      1 + number_of_replicas


      (documents_num) * (metadata_per_document + ID_size) * (no_of_copies)


      (documents_num) * (value_size) * (no_of_copies)


      total_dataset * (working_set_percentage)

      Cluster RAM quota required

      (total_metadata + working_set) * (1 + headroom) / (high_water_mark)

      Number of nodes

      Cluster RAM quota required / per_node_ram_quota

      Based on the above formula, these are the suggested sizing guidelines:

      Table 5. Suggested Sizing Guideline
      Variable Calculation


      = 1 for original and 1 for replica


      = 1,000,000 * (100 + 56) * (2) = 312,000,000 bytes


      = 1,000,000 * (10,000) * (2) = 20,000,000,000 bytes


      = 20,000,000,000 * (0.2) = 4,000,000,000 bytes

      Cluster RAM quota required

      = (312,000,000 + 4,000,000,000) * (1+0.25)/(0.85) = 6,341,176,470 bytes

      This tells you that the RAM requirement for the whole cluster is 7 GB. Note that this amount is in addition to the RAM requirements for the operating system and any other software that runs on the cluster nodes.

      CPU Overhead

      When sizing, you must account for raw CPU overhead when using a high number of buckets.

      • We recommend that you set aside 0.4 core/bucket for operational stability. This overhead does not account for any front-end workload which must be sized separately.

      • Monitoring is recommended for CPU usage and System Limits

      Sizing Index Service Nodes

      A node running the Index Service must be sized properly to create and maintain secondary indexes and to perform index scan for N1QL queries.

      Similarly to the nodes that run the Data Service, there is a set of questions you need to answer to take care of your application needs:

      1. What is the length of the document key?

      2. Which fields need to be indexed?

      3. Will you be using simple or compound indexes?

      4. What is the minimum, maximum, or average value size of the index field?

      5. How many indexes do you need?

      6. How many documents need to be indexed?

      7. How often do you want compaction to run?

      Answers to these questions can help you better understand the capacity requirement of your cluster, and provide a better estimation for sizing.

      The following is an example use-case for sizing disk:

      Table 6. Disk Sizes
      Input variable Value


      20 bytes

      Number of index fields


      Secondary index

      24 bytes

      Number of documents to be indexed


      When you calculate disk usage for the above test cases, there are a few factors you need to keep in mind:

      1. Compaction is disabled. This case illustrates the worst-case scenario for disk usage.

      2. Couchbase Server uses an append-only storage format. Therefore, actual disk usage will be larger than data size.

      3. Fragmentation will affect the disk usage. The larger the fragmentation, the more disk you will need.

      The above index consumes 6 GB of disk space.

      Sizing Query Service Nodes

      A node that runs the Query Service executes queries for your application needs.

      Since the Query Service doesn’t need to persist data to disk, there are very minimal resource requirements for disk space and disk I/O. You only need to consider CPU and memory.

      There are a few questions that will help size the cluster:

      1. What types of queries do you need to run?

      2. Do you need to run stale=ok or stale=false queries?

      3. Are the queries simple or complex (requiring JOINs, for example)?

      4. What are the throughput and latency requirements for your queries?

      Different queries have different resource requirements. A simple query might return results within milliseconds while a complex query may require several seconds.

      The following is an example use-case for sizing CPU:

      Assume that you have a user profile store, which stores a user’s name and email address. You would like to query based on a user’s email address, and you create a secondary index on email. Now you would like to run a query that looks like this:

      Select * from bucket where email = "foo@gmail.com"

      By default, N1QL uses stale=ok for a consistency model.

      It was observed that this query utilized 24 cores completely to achieve an 80% latency of 5ms against a bucket of 20M documents.