Cloud Engines
Cloud Engines are managed remote engines, with scale-out, distributed cache, access control and observability. Set --engine=cloud to run any pipeline on Dagger Cloud. Dagger streams local files, secrets, network tunnels and your terminal as needed, so you can run many agents without melting your machine.
- No infrastructure to operate. You do not manage Kubernetes, load balancing or cache volumes.
- One cache for everyone. Developers, agents and CI can safely pool compute and cache, so work done once is not done again.
Transparent usage pricing. Dagger bills the sum of each pipeline's exact resource usage, from open-source telemetry. Invoiced monthly with a minimum spend of $20 for individuals, $200 for teams.
- Pipeline CPU: 0.5¢ / CPU-min. Exact CPU time used by your pipeline, measured from Linux cgroups.
- Pipeline memory: 0.5¢ / GB-min. Actual memory used by your pipeline, sampled every second from Linux cgroups.
- Pipeline inputs: 2.5¢ / GB. Data that your pipelines pull in: git objects, container images, HTTP fetches, local file uploads, container ingress.
- Pipeline outputs: 10¢ / GB. Data that your pipelines push out: container images, container egress.
Idle capacity is free. Other providers bill the full machine for the full duration of each job, regardless of how much you actually use.
Repeat work is 5× cheaper. You pay 5× less for operations that could be served from cache, whether or not they actually are. This rewards you for optimizing your pipelines, and rewards us for accelerating them. Other providers earn less as your pipeline gets faster. An operation is repeat work if it has the same inputs as one that you ran in the last 30 days.
For example, one customer's measured usage in September costs $94 on Dagger with no repeat-work discount, against an estimated $197 on per-minute runners.
The Dagger Engine is open source. It runs on your local machine or your own server, at no cost.
$ dagger check --engine=cloud