Scaling with Atlassian Data Center

 

In this guide, we'll take you through how Data Center can help you address pain that comes with growth and increased demand. We'll also give you a brief overview of what to expect when evaluating whether Data Center is right for your organization.

 


 

Real impact

Recently, a longtime JIRA customer switched from a single server deployment to a clustered Data Center deployment to meet their scale needs as they grew. You can read more in their case study, but to give you a rough idea, they had:

  • 10,000 developers 
  • 1.5 million issues 
  • 3 million comments 
  • 1,500 projects. 

Using Data Center, the customer also architected their infrastructure to intelligently route REST API traffic to one specific node. It allowed them to alleviate the load at its application nodes. As their usage grew, they lowered response times while also significantly reducing the number of outages they experienced.

While part of this success story is unique to this customer, many organizations inevitably find themselves in a similar position of growth and demand. You can’t afford downtime or performance degradation during times of growth.

The next phase

Whether your organization consists of 500 users, 5,000 users, or 50,000 users, you may already be thinking about your next phase of growth, whether incremental or more substantial, and how it might impact application stability.

Growing pains 

It's important to remember that your total number of users is only one variable that can affect stability. Concurrent usage, custom configurations, add-ons, and API calls are all factors that can affect your application's performance .

Often, applications are initially configured with a smaller user count and less complexity in mind. As teams expand, and more business units across the organization adopt Atlassian tools, the increased load can start stressing those initial configurations. Additionally, with increased maturity, teams start to perform more complicated operations (implementing custom scripts, for example) that can further tax the system. This growth can significantly increase concurrent usage, stretching those initial configurations, particularly in a single server environment, and cause application instability, longer response times, and other performance degradation.

Things also become more complicated if other development teams provision separate Atlassian applications, either to accommodate short projects or because they are unaware of existing applications. Once you start integrating and collaborating across these applications, you'll find yourself managing a federation of instances, potentially running different versions of the application, and add-ons. The task of managing, upgrading, and maintaining integrations between each of these servers and add-ons individually can be very challenging.

In addition, you may need to evaluate other aspects of your systems to ensure they fully meet the needs of an enterprise business. Avoiding downtime becomes a major priority as even small outages can negatively impact productivity and revenue. As your teams' tools become mission critical, you'll spend considerable effort minimizing the risk of downtime as much as possible, especially if your users are located all over the world. A surge in traffic, running computationally intensive operations, a hardware failure, application upgrades, or manually adding another application instance to your environment could all result in downtime. Good governance for all your IT systems, including access management and disaster recovery plans, is essential.

Whether you are currently growing or planning for the future, you need more powerful control over your applications to manage these issues.

Atlassian Data Center

Data Center is a self-hosted deployment option that gives you more flexibility and control over provisioning and maintaining your Atlassian applications. Data Center applications (JIRA Software, Confluence, Bitbucket, and JIRA Service Desk) provide all the same functionality as their single server counterparts, but are deployed with a load balancer and a shared filesystem to run in a multi-node cluster. There are also Data Center only features which have been specifically designed for the enterprise environment.

Here's what a typical Server and Data Center deployment might look like: 

 

 

By deploying Data Center you get:

Horizontal scaling

The ability to grow your cluster horizontally, by adding more nodes to the cluster, allows for more concurrent users, as the load balancer routes incoming traffic to the nodes in the cluster according to load they're experiencing. This load balancing is the most immediate gain from running your application in Data Center. Adding a node to your cluster for additional capacity is simple and does not cause downtime as you are just adding another copy of the application to the load balancer. Data Center handles all required syncing processes.You can see the effect this has on your system's capacity for concurrent usage in the test results we've published for JIRA SoftwareBitbucket, and Confluence Data Center.

Cluster management

You can further leverage other benefits of a multi-node cluster and tailor them specifically to your environment. With Data Center, you have full control over how you segment traffic across your application nodes to ensure seamless access to all your users. Evenly balance traffic across your nodes or send targeted traffic to a particular node. The multi-node configuration allows for High Availability, so that if a node fails, the load balancer sends traffic to other running nodes.  Data Center's Disaster Recovery capabilities makes recovering from a complete system outage easier. With Data Center, you can run a "cold standby" copy of your system, completly separate from your production environment. By monitoring your cluster, can measure and tweak the performance of your system to better meet the needs of your users as your organization continues to evolve.

Enterprise features

Data Center currently offers, and will continue to introduce, enterprise specific features across its applications. For example, Bitbucket Data Center offers Smart Mirroring which allows developers in remote locations to access code faster. JIRA Data Center allows zero-downtime upgrades, and all Data Center applications offer SAML single sign-on, which helps you manage users across all your Data Center applications using your preferred identity provider. 

Atlassian Enterprise Services

We're here to help! Data Center provides the foundational tools enabling you to create and maintain a high-performance system tailored to the needs of your organization. Our Enterprise PartnersPremier Support staff, and Technical Account Managers are ready to help you through the entire end to end process from concept to production. These teams are committed to helping you build an enterprise-grade system tailored to your organization's needs.

Evaluating Data Center

Chances are you're already using Atlassian products. We offer trial Data Center licenses so you can evaluate how Data Center performs in your environment.

This high-level overview will help you identify how Data Center will best fit your environment and where you can make further optimizations to fine tune performance.

Before you begin

  • Investigate current installation for any possible performance optimizations
  • Upgrade applications
  • Verify add-on compatibility with Data Center
  • Acquire any additional infrastructure required
  • Benchmark current system metrics
  • Develop a scorecard of goals and metrics to evaluate your Data Center implementation against

Implementation

  • Deploy a basic cluster configuration in a test environment
  • Ensure nodes are functional, in sync and test basic application functionality
  • Invite a test batch of users to measure system performance metrics
  • Compare metrics against your original benchmarks

This process is a crucial investment to make sure Data Center meets your organization's needs at scale.

Amazon Web Services

If your target infrastructure runs Amazon Web Services, you can make use of our existing CloudFormation templates to get up and running. Here's some existing documentation to get you started:

Path to success

Many of our customers operating successful Data Center deployments have all taken a methodical approach to analyzing, evaluating, and testing to ensure that Data Center specifically meets their organizational needs. A key driver of this success is the testing phase. Involved testing will provide you with data to better manage expectations and predict how well your system will function during day-to-day operations, growth, and incidents. Some customers take a few months to test thoroughly and evaluate Data Center, and other customers take a bit longer. Wherever you fall on the spectrum, it's an important aspect to fully understand how Data Center will meet the needs of your enterprise organization.

Atlassian Data Center has helped nearly 1,000 customers meet the needs of their users and business while scaling. We believe that it can become an important part of your infrastructure strategy as well.



 




Last modified on Aug 24, 2017

Was this helpful?

Yes
No
Provide feedback about this article

Not finding the help you need?

Ask the community

Powered by Confluence and Scroll Viewport.