Running Meteor

Once you have a recipe (or a folder of them), meteor run executes the extraction.

With the Meteor binary

# run a single recipe
meteor run recipe.yaml

# run all recipes in a directory (in parallel, not recursive)
meteor run recipes/

# run all recipes in the current directory
meteor run .

Preview without sending data

Use --dry-run to run the extractor and processors without pushing anything to sinks. Combine it with --limit to stop after a few records:

meteor run recipe.yaml --dry-run --limit 10

This is the safest way to check a new recipe or debug a config.

With Docker

docker run --rm \
  -v $(pwd)/recipes:/recipes \
  raystack/meteor meteor run /recipes/

Mount your recipes into the container and pass credentials as environment variables with -e METEOR_....

After the run

Meteor prints a summary per recipe: success or failure, duration, record count, and a per-entity-type breakdown. A directory run reports every failure at the end and exits non-zero if any recipe failed. See Architecture for the full error-handling rules, and Deployment for running Meteor on a schedule.