Rippling: Re-building Async processing using Temporal

maksim-bezsaznyj

Maksim Bezsaznyj

Rippling

malay-kurwa

Malay Kurwa

Rippling

Need a closer look? Download and review slides presented at Replay 2023 here.

At Rippling, we have built a HR Product loved by people building and running their companies, and people love Rippling for ease of defining their tasks and forgetting, and our tools do their jobs. When Rippling was built, we chose simple technologies, to prove the product, quickly iterate over changes and new features.

As we scale our product to handle the same use cases, we run into a few of the challenges such as automation for our jobs. We have an inhouse jobs execution environment called ETA (Estimate Time of Arrival), which gets most of our work done through a tight integration with Django via custom implementation hooks. At our current scale, we see wear and tear - owing to maintenance challenges, like missed executions, non-stateful retries, observability. We find Temporal to be a right fit here on our use-cases. As one test of Temporal, we have onboarded Filling Factory and have gained a few insights.

As Rippling scales, we are moving towards an asynchronous processing paradigm, and Temporal will play a crucial partner in our success.