Serverless Workers

7 posts · Feb 2026 – Sep 2026

What this trail covers

Serverless Workers enable Temporal Workflows to run on serverless compute platforms like AWS Lambda and GCP Cloud Run, scaling automatically based on workload. This post details the integration process with AWS Lambda and Amazon Bedrock AgentCore Runtime, focusing on how Temporal invokes, scales, and gracefully shuts down Workers on behalf of the user. It highlights automatic state management for the Worker lifecycle, routing tasks through Temporal Activities, and leveraging AgentCore's scaling capabilities. This reduces operational overhead and allows for elastic compute for agentic applications.

All Temporal trails

Sep 2026Run durable AI agents on Amazon Bedrock AgentCore with Temporal Serverless WorkersThis post introduces the prerelease of Amazon Bedrock AgentCore Runtime as a compute provider for Temporal Serverless Workers. It details a reference architecture where Temporal Workflows act as the durable agent loop, coordinating model calls and tool use as Temporal Activities. AgentCore Runtime hosts the Temporal Worker, and Temporal Serverless Workers manage capacity by invoking AgentCore when tasks are ready. The integration leverages Temporal's Strands integration to turn model and tool operations into durable Activities, and AgentCore's microVM pricing for cost-effective compute during idle periods.
Sep 2026From dedicated Workers to Serverless Workers: Migrating Temporal Workflows to AWS LambdaThis post details the migration of Temporal Workflows from dedicated Workers on Kubernetes to Serverless Workers on AWS Lambda using the Python SDK. It covers the architecture of Serverless Workers driven by a Worker Controller Instance (WCI) that reacts to sync match failures and task queue backlogs. Key migration steps include adding a Lambda handler using `temporalio.contrib.aws.lambda_worker`, configuring Worker Versioning (PINNED or AUTO_UPGRADE), tuning shutdown timings for long-running Activities (graceful_shutdown_timeout, shutdown_deadline_buffer, and Lambda invocation timeout), packaging dependencies with consideration for cross-platform builds (e.g., macOS to Linux), and trimming packages to fit Lambda's size limits. It also touches on the independent nature of each Lambda invocation and the lack of shared state or connection reuse.
Aug 2026Durable Digest: August 2026This post announces the Public Preview of Serverless Workers for AWS Lambda. This feature provides zero infrastructure management, instant autoscaling, scale to zero, built-in observability, and simplified versioning. Temporal Cloud customers can now upload Worker code to a Lambda function, and Temporal handles invocation automatically. When there's no work, functions scale to zero, and no billing is incurred. This builds upon the existing Serverless Workers capability by providing a specific, production-ready implementation for AWS Lambda.
Aug 2026Better together: The durability layer for your AWS agentic stackThis post details the integration of Temporal's Serverless Workers with AWS Strands Agents SDK and AWS Lambda. It explains how Temporal adds a durability layer to AWS's agentic AI stack by automatically managing state for the Strands agent loop, routing model invocations and tool calls through Temporal Activities. This ensures that agent execution can survive failures and resume from the last step using Workflow history. The post also discusses the use of Temporal's Codec Server for client-side encryption when using Temporal Cloud with AWS.
Jul 2026Durable Digest: July 2026Announces the pre-release of Serverless Workers for GCP Cloud Run, enabling Temporal Cloud to automatically manage worker invocation, scaling, and shutdown based on workload metrics. This extends the existing Serverless Workers capability.
Advertisement
May 2026Announcing new Temporal capabilities from Replay 2026This post announces the pre-release availability of Serverless Workers, which allow Temporal Workers to run on serverless compute platforms like AWS Lambda. It highlights that Temporal Cloud will automatically invoke, scale, and gracefully shut down these workers based on workload volume and metrics, eliminating the need for manual infrastructure planning and autoscaling strategy design. The goal is to reduce time to get started and improve cost efficiency by scaling to zero when appropriate.
May 2026Introducing Temporal Serverless Workers: Deploy Temporal Workers to AWS LambdaThis post introduces Temporal Serverless Workers, allowing Temporal Worker code to be deployed directly to AWS Lambda. It details the three-step setup process: uploading code to Lambda, creating an IAM role for invocation, and registering the Lambda function with Temporal. The post explains the invocation model where Temporal detects tasks, evaluates metrics, and invokes Lambda functions, which then start a Worker to process tasks before shutting down. It also covers the developer experience with a Go example and discusses use cases for bursty/event-driven workloads, low/intermittent volume, serverless-native organizations, and rapid startup. It contrasts this with traditional self-managed workers for sustained high-throughput workloads. Future work includes support for Google Cloud Run and additional SDKs.
Feb 2026Distributed privilege access with TemporalThis post introduces the integration of Serverless Workflows (CNCF specification) with Temporal for distributed privilege access management. It details how Serverless Workflows can be used to define declarative workflows that span multiple environments (local device, SaaS, AWS). The post explains how Temporal Task Queues are used to route specific activities to dedicated Workers in different environments. It then introduces Thand, an open-source platform that extends Serverless Workflows with Temporal capabilities, including Nexus. Thand uses 'providers' (e.g., 'thand: agent', 'thand: aws') within the Serverless Workflow definition to declaratively route tasks to appropriate Temporal Workers and environments, abstracting away direct Task Queue management. This enables distributed, reliable provisioning and de-provisioning of access and resources.