Durable Digest: July highlights

AUTHORS

Temporal Technologies

PUBLISHED

Jul 30, 2026

Recently shipped#

Serverless Workers for GCP Cloud Run is in pre-release. With Serverless Workers, you can run Temporal Workers on serverless compute. Reduce the time to get started as Temporal Cloud automatically invokes, scales, and gracefully shuts down (scaling to zero, if appropriate) Temporal Workers on your behalf based on workload volume and metrics.
LangGraph (Python) and LangSmith (Python and TypeScript) integrations are now in Public Preview. We believe there is a difference between an agent framework and an agent runtime. While LangGraph tries to be both, Temporal’s position is clear. These integrations are for customers and developers who want to keep LangGraph but standardize production execution on Temporal.
Billing API and Billable Action Metrics for Temporal Cloud are now Generally Available. Temporal Cloud integrations with Datadog and Vantage are also Generally Available, helping customers track spend, optimize costs, and identify usage trends. Three supporting blog posts can help customers get started.
Worker Status UI Update: Workers that support Worker Heartbeating are now visible in the UI with a new first-class Workers section. Users can view heartbeating Workers, see Deployments alongside Workers, inspect Workers tied to a Workflow’s Task Queue, and drill into Worker details such as CPU, cache hit ratio, and available Task slot.

Builder spotlight#

This month we’re featuring Jeff Romine, who built a Temporal E-commerce Demo that’s featured on our Code Exchange. Jeff describes this demo as a “A full-stack e-commerce application built entirely on Temporal durable execution. Every state transition — from adding an item to a cart through order fulfillment and delivery — is a Temporal workflow. No message queues, no cron jobs, no saga orchestrators. The business logic is the infrastructure.”

Jeff skillfully used Temporal's primitives as design patterns to build this demo. Workflows run as testable state machines, inventory follows a CQRS read/write split, and parseable Workflow IDs plus Search Attributes let one query trace an entire customer journey. Built on the Temporal TypeScript SDK with Next.js and Cassandra, it's a clear, hands-on blueprint for what durable execution looks like taken all the way down to the foundation.

Check out his Github to learn more!

How to Temporal#

Validated Patterns are practical, field-tested implementation guides that help you solve complex integration concerns and business problems that may arrive when building on Temporal. Each pattern provides prescriptive, field-tested guidance for building production-ready applications with Temporal, so you can spend less time evaluating approaches and more time shipping. Whether you're designing a new system or scaling an existing one, explore the growing library of proven patterns to accelerate your next project.

Temporal Nexus in .NET (Preview): What It Is, When to Use It, and a Practical Walkthrough covers the new Nexus feature from a .NET perspective, which is still underserved in the community compared to Go and TypeScript. Rebecca explains what Nexus is, the scenarios where it makes sense, and walks through a working example. If you're building on the .NET SDK and curious about cross-namespace communication, start here.

Migrating from ARQ to Temporal for Agentic Workflows is part one of a two-part series from Marc Novakowski from Magnitude on swapping out ARQ for Temporal to run agentic workflows. If you've been using a lightweight Task Queue and started hitting its limits as your agents get more complex, this is a practical first-hand account of why and how to make the switch. Part two is on the way.

Questions from the field#

In the #python-sdk channel, Ashu asked whether the approach of using decorators to bind a request ID ContextVar into Temporal's Activity and Workflow loggers was the right pattern, or whether interceptors were the more idiomatic path. Maxim Fateev suggested passing the request ID via headers injected by interceptors, while Tim Conley noted that the SDK's loggers don't have deep extensibility and that interceptors were the cleaner approach. Basit Raza provided a detailed walkthrough of implementing WorkflowInboundInterceptor and ActivityInboundInterceptor to centralize request ID binding, pointing to the context propagation sample in temporalio/samples-python. Ashu confirmed that the interceptor approach worked.

In the #operations channel, Phillyp described a persistent memory growth problem on temporal-history pods running server 1.30.4, where memory climbed from 500 MB to 2 GB over two days despite byte-based cache limits and background eviction being configured. Tom Wheeler pointed to a related Discourse thread where an earlier reporter found relief by using a count-based limit for mutable state cache alongside a size-based limit for the events cache, and by setting GOMEMLIMIT on History Service pods. Community member Peerat Vichivanives confirmed that the same fix worked on version 1.30 as well. Phillyp marked the thread resolved.

Join us live#

Whether you prefer to join online or in person, we've got plenty of opportunities to connect this month. Tune into our Vibe Check livestream series for conversations with industry leaders, register for our upcoming webinars, or meet the Temporal team.