Local Mode

Debug and iterate faster

You can get started in Conducto’s free cloud with minimal effort.

  • For CI/CD, install the GitHub integration and immediately start running pipelines triggered by GitHub events.
  • For data science, pip install conducto then launch cloud pipelines from your local machine.

Once you have successfully launched a few cloud pipelines and are ready to fully embrace Conducto, you can unlock its full power with free local mode. This requires a little bit of setup on your machine.

Reproduce and debug errors instantly

Select an errored node in the browser and click Debug to get a shell command.

Pick a node to debug

Run it in the terminal to start a Docker container with the node’s code and environment, reproducing the error interactively.

Explore with a shell

You actually have two debug options. Snapshot is isolated from your system, and any changes you make go away once you exit the container. Live Code mounts your local code into the container, so you can develop as normal and test changes instantly.

Use uncommitted code for faster iteration

You don’t need to commit debugging code to Git. Conducto can run pipelines with your local, uncommitted code for straightforward debugging and prototyping.

from conducto import Image

# Baseline: include all code in a public Git repo or any
# private repo with Conducto's GitHub integration installed.
Image(copy_url="https://github.com/conducto/examples")

# Get code from a private repo with a personal access token.
Image(copy_url=f"https://{user}:{token}@github.com/{owner}/{repo}")

# Enabled by local mode: Include all code in this file’s directory
Image(copy_dir=".")

For more about this, see Debugging.

Launch free local pipelines

All users get 1 CPU of free cloud workers, but pipelines launched with --local are only limited by the resources on your machine. Even after you upgrade to paid cloud mode for greater scale, it is still useful to prototype pipelines in local mode. Local pipelines start faster because there are no cloud latencies when everything is on your own machine.

Conducto is built on containers, so pipelines run the same on your machine as in the cloud. Once you are done prototyping, just replace --local with --cloud to run at scale.

For more about --local and --cloud, see Local vs Cloud.

Chat with us for a live demo right now!
(If we're awake 😴)

avatar