March 2022 Summaries

OpenTelemetry is a vendor-agnostic standard for observability that aims to provide metrics, logs, and traces across implemented languages. It separates instrumentation from downstream storage and query layers, allowing users to decouple them. This enables the use of multiple open-source projects with different instrumentation APIs and ecosystems, simplifying observability data collection and analysis. OpenTelemetry provides a standard API for instrumentation and language SDKs that generate observability data, making it easier to implement and manage observability in applications. The library also supports auto-instrumentation, which automatically generates spans and attributes without requiring code changes, providing best practices for frameworks and removing downstream concerns from observability generation. With OpenTelemetry, users can mix manual and automatic instrumentation as needed, combining them to suit their specific use cases. Additionally, OpenTelemetry provides support for exceptions, including error events with stack traces, making it easier to diagnose issues in distributed systems. Overall, OpenTelemetry is a powerful tool for implementing observability in applications, providing flexibility, scalability, and ease of use.

We invited Ken Ahrens, co-founder and CEO of Speedscale, to share how their company modernizes and automates testing practices for cloud infrastructure by providing API traffic data going into and out various microservices. Thanks to TimescaleDB, Speedscale's UI loads quickly and provides the ability to drill down into high-fidelity data. Speedscale is one of the first commercial technologies utilizing actual API traffic in order to generate tests and mocks, helping Kubernetes engineering teams validate how new code will perform under production-like workload conditions. The platform provides unparalleled visibility, collects and replays API traffic, introduces chaos, and measures the golden signals of latency, throughput, saturation, and errors before the code is released. Speedscale's Traffic Replay is a modern load, integration, and chaos testing framework that can be used by organizations providing critical SaaS products or teams responsible for performant infrastructure to generate quality automation quickly without manual scripting. The platform enables engineers to isolate and contract/performance test smaller components in tightly coupled architectures, allowing rapid testing iterations and faster development cycles. By leveraging traffic to automatically generate sophisticated mocks, Speedscale's leadership team comes from companies like New Relic, Observe Inc, Wily Introscope (bought by CA Technologies), and iTKO (bought by CA Technologies). The company uses a transparent proxy to capture API transaction data, streams it to AWS S3, and filters for test and mock scenario creation. Speedscale's architecture is built around TimescaleDB, which enables fast query performance and cost savings, with the platform reducing AWS cloud costs by 35%.