Serverless Workers for Cloud Run

6 posts · Jan 2025 – Aug 2026

What this trail covers

Serverless Workers enable Temporal Workflows to run on serverless compute platforms like AWS Lambda and Google Cloud Run, scaling automatically based on workload. This post details the integration with Google Cloud Run Worker Pools, providing a custom, highly responsive, and efficient autoscaling algorithm that allows users to skip manual setup and tuning. Temporal Cloud now manages the autoscaling of Worker Pools based on key metrics indicating work queueing up, scaling down to zero when appropriate. This update focuses on leveraging Worker Pools and CREMA for enhanced autoscaling based on Temporal Task Queue backlog.

All Temporal trails

Aug 2026Durable Digest: August 2026This post announces the Public Preview of Serverless Workers for AWS Lambda and Google Cloud Run, enabling zero infrastructure management and auto-scaling. It also highlights the Public Preview of the Google GenAI integration and the Pre-Release of the Deep Agents integration, both of which provide Temporal's durability to AI applications running on serverless platforms. Additionally, it introduces Projects for Temporal Cloud (Pre-Release) for resource organization, Standalone Nexus Operations (Pre-Release) for direct invocation, Activity Operations (Public Preview) for granular control over Activities, and Workflow Pause (Pre-Release) for investigating broken Workflows without termination. Worker Performance (Generally Available) with Poller Autoscaling is also highlighted.
Jul 2026Serverless Workers for Google Cloud RunThis post details the integration of Temporal Serverless Workers with Google Cloud Run Worker Pools. It explains how Temporal monitors key metrics to dynamically adjust capacity within Worker Pools, eliminating the need for manual autoscaling setup and tuning. The post highlights that worker code and activities remain the same, with deployment to Cloud Run involving Docker container creation and Worker Pool specification. Temporal then manages scaling activities, including scaling down to zero.
Feb 2026Distributed privilege access with TemporalThis post introduces Thand, an open-source implementation of the Serverless Workflows specification that adds extensions for Temporal capabilities, including Nexus. Thand enables distributed provisioning and access management by allowing Serverless Workflows to be routed to specific Temporal Workers and environments using 'providers' and custom Task Queues. It demonstrates how to use Serverless Workflows with Temporal for tasks like checking entitlements via gRPC, running Osquery on local devices, and provisioning AWS access, all while abstracting away the complexity of Task Queue management and Worker topology.
Apr 2025Announcing KEDA-based auto-scaling for Temporal WorkersThis post announces the integration of KEDA (Kubernetes-based Event-Driven Autoscaling) with Temporal Workers. It details how KEDA's Temporal scaler can dynamically scale the number of Temporal worker pods based on the size of the Task Queue backlog. This provides benefits such as cost efficiency (scaling to zero), high throughput during traffic spikes, and operational simplicity. A demo application using the Order Management System (OMS) reference app is provided to showcase this functionality, visualizing Task Queue backlogs and worker pod scaling in real-time.
Jan 2025Temporal Workers & Google Cloud RunThis post updates guidance on deploying Temporal Workers to Google Cloud Run by leveraging the new Worker Pools resource type and CREMA (Cloud Run External Metrics Autoscaling). It details the configuration of Worker Pools, the use of a sidecar container with OpenTelemetry Collector for Prometheus metrics, and the configuration of CREMA for autoscaling based on Temporal Task Queue backlog depth. The scaling formula and key tuning parameters for CREMA are explained, including `targetQueueSize` and `activationTargetQueueSize`.
Advertisement