March 2024 Summaries

Flightcontrol is a Platform-as-a-Service (PaaS) that specializes in deploying production applications with a focus on reliability, flexibility, and performance, leveraging AWS infrastructure. The company has announced the general availability of its Fargate scheduler job runner, designed to efficiently handle scheduled and on-demand tasks by automatically deploying and shutting down dedicated servers to run jobs, ensuring cost-effectiveness by only charging for execution time. This service is particularly beneficial for long-running, resource-intensive tasks or when a dedicated worker server is unnecessary. The scheduler offers ease of use with a user-friendly interface that includes intuitive cron scheduling, timezone management, execution history visibility, and the ability to pause jobs. Users can customize the resource allocation per job, and the platform supports various build types, including Nixpacks, custom Dockerfiles, and image registries. Additionally, it provides comprehensive logging and notifications for job executions, with options for advanced monitoring through sidecar containers that integrate with tools like Datadog.

Temporal is a robust open-source workflow orchestrator initially developed at Uber, utilized by Flightcontrol for backend workflows like infrastructure provisioning and deployment pipelines. Temporal's error handling is intricate, necessitating a custom system for managing complex errors, which the author details in their post. The two primary concepts in Temporal are Workflows and Activities, where Workflows handle non-async logic and Activities manage async logic. Errors in Temporal, due to its distributed, event-driven nature, are serialized and retried by default unless specified otherwise, requiring careful handling of custom errors. The author describes using Temporal Interceptors, both Activity and Workflow types, to customize error handling and ensure the transmission of custom error data. These interceptors convert CustomErrors to non-retryable ApplicationFailures and unwrap errors to maintain the original error context. The post provides insights into meeting error handling requirements across different scenarios, demonstrating how to integrate custom error management within Temporal's framework effectively.