● Covered by 1 source · 1 reportLow impact1 neutral
DoorDash's Flux Platform Automates 130,000 Engineering Tasks Monthly in the Cloud
Key points
- Flux platform handles 130,000 engineering tasks monthly.
- Moved agent workloads from laptops to cloud for improved security and performance.
- Supports over 25,000 automated code reviews weekly.
- Uses Firecracker microVMs for isolated execution sandboxes.
Migration to Cloud-Based Automation
DoorDash has transitioned its engineering agent workloads from individual developer laptops to its proprietary cloud platform, Flux. This move has enabled the platform to automate a significant volume of tasks, reaching 130,000 engineering tasks in one month during 2026. The Flux platform also manages over 25,000 automated code reviews each week, utilizing more than 300 playbooks for over 10,000 weekly invocations.
Addressing Limitations of Local Execution
The development of Flux was prompted by limitations encountered with agent workloads running directly on developers' laptops. Local execution presented issues such as restricted CPU and memory resources, dependency on the developer's device connectivity, and potential security risks by granting autonomous agents access to credentials and internal systems. Furthermore, monitoring agent activity, system access, and operational context was challenging with local execution.
Architectural Components of Flux
Flux is structured around four core primitives: cloud sandboxes, an MCP gateway, reusable playbooks, and invocation surfaces. Playbooks define the specific tasks, while cloud sandboxes provide isolated execution environments. The MCP gateway controls access to internal systems, and invocation surfaces allow workflows to be initiated through various channels including Slack, GitHub, cron jobs, command line interfaces, or conversational interfaces.
Isolated and Secure Execution
The sandboxes within Flux leverage Firecracker micro virtual machines to ensure isolated execution of agent workloads. Each environment is provisioned with necessary repositories, development tools, secrets, and runtime dependencies for its task. DoorDash reports a 95th percentile service level objective of under five seconds for end-to-end sandbox setup, which includes microVM startup, repository cloning, tool installation, and agent harness configuration.
Agents access DoorDash's internal systems via Agent Gateway, an in-house MCP gateway that enforces scoped permissions and logs all agent activity for auditing and policy enforcement. Playbooks, defined in YAML, specify tasks, required tools, permissions, validation, and safety boundaries, allowing for a combination of agent-driven steps and deterministic code.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
DoorDash has migrated its engineering agent workloads from developers' laptops to its cloud-based Flux platform, which processed 130,000 engineering tasks in a single month in 2026. This shift addresses limitations of local execution, such as CPU/memory constraints and security concerns, by providing isolated, monitored environments for automated workflows.