Hi people!
I need to run a stress test, but to call the HTTP endpoint I want, I require data that isn’t ready initially. Here’s how it works: The load test (LT) I’m creating calls an endpoint to generate EntityA
, but creating EntityA
requires the ID of EntityB
, and its instances are being created and published to an SQS queue by another continuos flow. I’m considering a design where K6 consumes messages from this SQS queue, captures a EntityB
id
, and calls the target endpoint to generate load. However, I’m concerned about this approach—I haven’t seen any load test consume data while simultaneously calling endpoints to generate load. The official K6 SQS library doesn’t even support message consumption.
Hi @lucaslgr,
Welcome to our community forums!
I’m not sure about the entire workflow either, as I have never personally worked with a scenario like this. But I’d encourage you to give it a try, and let us know about your results, or even raise specific issues/concerns if any.
Generally speaking, I know that some user journeys involve interaction with multiple endpoints and/or message queues, and so a simulation of a certain amount of such user journeys concurrently requires dealing with these interactions, unless you want to define a more granular scenario.
The official K6 SQS library doesn’t even support message consumption.
Have you seen seen the support for SQSClient.receiveMessages(queueUrl, [options])
?
Isn’t it enough to consume the messages?
Bear in mind that while do offer k6-jslib-aws to make it easier to define workloads that interact with AWS resources, you can always contribute with any missing operation/service, and/or create your own library.
I hope that helps!
Thanks @joanlopez I’ll give it a try