Durable Execution for Unreliable Data

Imaxxing solves the problem of monitoring theater seat availability across the country by using Temporal to manage complex, long-running workflows. The core challenge in this project was the instability of downstream data sources, which lacked modern uptime guarantees. By implementing Temporal's "durable execution" model, the application ensures that code transitions are persisted at every key point. If a network blip or service failure occurs, the workflow can resume from the exact point of failure rather than crashing or losing state, providing an "at-least-once" execution guarantee that is critical for real-time alerting.

Entity Workflows and Intelligent Batching

The architecture relies on "entity workflows" to map business logic to specific entities. Each user subscription is a distinct workflow that tracks user preferences (theaters, party size). A separate monitoring workflow tracks theater seat maps. When new seats are detected, the monitoring workflow sends a signal to the relevant subscription workflows.

To prevent notification fatigue, the system implements a 60-second "cooldown" timer within the subscription workflow. When a seat opens, the workflow waits for one minute to collect any additional notifications before sending a single, consolidated digest to the user. This approach optimizes infrastructure efficiency by allowing the worker to sleep while waiting, freeing up resources for other tasks.

Serverless Scaling with Cloud Run

By deploying Temporal workers on Cloud Run, the project achieves autoscaling tied directly to Temporal task queues rather than secondary metrics like CPU usage. This solves the common problem of over-provisioning worker pools to handle traffic spikes. When a popular movie triggers a surge in notifications, the serverless workers scale up instantly to meet the demand and scale down when the queue clears, keeping costs to a few hundred dollars per month despite processing thousands of active subscriptions.