Speedscale is seeking to cut time and errors out of the Kubernetes and container delivery pipeline with their ability to discover API connections, automatically generate tests and data, replay traffic, and spin up realistic lab environments and reports within the tight time windows of cloud-native development.
As it turns out, procedurally testing the happy paths with unit tests when releasing tens or hundreds of Kubernetes pods a day is entirely insufficient for ensuring system resiliency at scale. Candidate microservices deployments need to be exercised in the context of a shifting array of upstream and downstream dependencies, but there’s seldom time to wait for that.
Speedscale’s Traffic Replay platform gathers the traffic of API calls and responses, and as an operator in the cluster, it constructs a sort of ‘instant test sandwich’ by surrounding the candidate service with an ephemeral K8s environment.
By exercising the app from both ends with thousands of auto-generated tests containing obfuscated but valid test data, CI/CD automation tools can get back realistic observability data such as the ‘golden signals’ of latency, throughput, saturation and success rate prior to each release.
The company also just announced a new Traffic Viewer which gives developers and SREs a diff-style scope into traffic that helps users drill down into exceptions and decide where to prioritize additional testing.
©2021 Intellyx LLC. At the time of writing, Speedscale is not an Intellyx customer, and Jason English is an advisor to the company.