Why performance is decisive

“You can't control what you can't measure": slow response times, sluggish processing of requests and time-outs in a project workflow require rapid optimization of system performance. In order to be able to measure the performance of an Atlassian toolchain, individual tests must be carried out. Background: Every environment and system usage is unique. catworkx relies on the pairing of open source software and its own internal tools, such as catworkx SPIN (stress app), when controlling performance values. This allows behavioral information of a specific system to be accurately mapped over a defined time frame, while maximizing the stress level of the Atlassian instance.

This realization was the catalyst for catworkx to invest in measuring the performance of the Atlassian toolchain. As centralized processes and business functions in medium and large organizations increase the need for workflow management, documentation, collaboration and compliance policy implementation, all parts of the toolchain need to be scrutinized.

In the following example, the catworkx team took care of a Jira system and put together a tool palette for it, with which business customers can clearly visualize the essential information. This allows business managers and IT staff to understand and identify bottlenecks or stumbling blocks that hinder service. catworkx has applied its skills in improving the performance and increasing the system stability of the Atlassian toolchain many times. This knowledge and subsequent set-up is the foundation of how we help today's clients get rid of yesterday's problems and prepare for tomorrow's business needs.

Well aware that every environment and every system usage is unique and requires individual investigation, the starting point for each individual system must be scrutinized. After evaluating several tools for our needs, catworkx came across a tool set consisting of Gatling (open source testing framework), InfluxDB (open source TSDB database for storing large amounts of data from time measurement series) and Grafana (open source metrics dashboard), which meets our requirements in terms of scalability and practicability.

These external tools are bundled with our own internal tools, such as catworkx SPIN (stress app), to collect behavioral information of a specific system - while maximizing the stress level of the Atlassian instance - over a defined period of time.

The following example is about a customer system (Jira) that has attracted attention due to slow response times, sluggish processing of requests and time-outs. After the first cycle, we found out that every single change to the tool's setup or its configuration needs to be cross-tested to investigate its benefits. Changing more than one condition at a time did not prove to be the right approach, as overlaps and side effects can negatively influence the measurements.

These steps increased the overall system performance and fluidity to an acceptable level, allowing the customer to continue using their system with the optimized parameters. The measures achieved an effect of 30 to 60 percent in individual aspects. Continuous monitoring ensured that overlapping side effects were removed from consideration. The result was a satisfied customer who did not have to buy a new (larger) system to keep pace with their business requirements.