Luckily, you don’t need to wait and react to issues like stalled integrations and carrier manifest close errors. By adapting a concept called “chaos engineering”, which was originally implemented at firms like Netflix, you can proactively stress test your systems no matter how small your IT team is.
In short, chaos engineering is the idea of using automated tools to simulate potential shocks like data volumes, network latency, and server failures. Depending on your deployment model (on-premise vs cloud), you may not be susceptible to all possible shocks, but challenges like latency and sudden volume spikes are relevant to execution systems regardless of your server location.
Of course, the needs of a video streaming platform are very different from a transaction-processing system, but this practice could avoid very painful days in an operation later. Translated to warehouse “chaos,” you may want to consistently test all code rollouts for their handling of unusually large integration payloads, suboptimal wi-fi conditions, confused user interactions, and unicode order data.
Over time, configuration changes and logic adaptations can lose touch with the original solution summary. As the original implementation resources move on to other assignments, it is hard to remember when, for example, a 3PL’s largest client indicates a hot order using an integrated field, which must then trigger custom packing notes. Is this “chaos”? Perhaps not, but when the client cares about these orders more than any other, it could result in chaos for the business if these orders are mishandled.
So how can you implement chaos engineering in a way that doesn’t risk production? With some creative application of test automation, like the Cycle® platform.
Unless you have a background in ERP, CRM, or other back office systems, you may have never encountered test automation in your implementations before. Test automation enables you to automate system and user-driven interactions with applications, reducing the manual effort required to complete functional and regression testing in your development processes.
Consumer-facing and back-office apps began adopting test automation years ago due to the business criticality and IT resource availability. Vendors like Smartbear and Tricentis led the way with model-based approaches that use intelligent automated analysis to build a model of a system’s interface and data flow to accelerate test scenario creation after technical delivery.
Cycle® was created by warehouse system implementers to address the unique needs of execution systems directly – particularly the tendency of requirements to evolve over the course of projects and for development project phases to last until near “go-live”. These two realities of execution system implementations compress testing timelines and increase the risk of quality challenges. Execution systems also tend to have processes that branch in complicated ways based on client requirements, which makes it difficult to model the complete business process. Finally, execution system tests need to support seamless interaction across interfaces like web, automation, APIs, and databases.
Since Cycle® test automation enables you to start testing earlier (shift left) in your execution system implementation, it also helps you document and test the what-if scenarios that can cause chaos months later if not accounted for.
Interested in learning more about test automation or how Cycle® may benefit your organization? Contact us to continue the conversation.