convalesce
/changelogPlugins

What changed in the plugins.

Release notes for the Convalesce plugins that run beside your pipeline: Airflow, Dagster, Prefect, Great Expectations and the Spark listener. Each Python plugin is one pip install; Spark is two lines of config.

Packages

  1. 1.0.0

    • Send the last lines of a failed task's output.

    Release 1.0.0 on GitHub

  2. 0.3.0

    • Send a run's settings, with credentials as keyed hashes.

    Release 0.3.0 on GitHub

  3. 0.2.1

    • Run on Glue, Databricks serverless and YARN as they are, and close the gaps real runs found.

    Release 0.2.1 on GitHub

  4. 0.2.0

    • Retries: carry out retries with the key and sign-in a deployment already has.
    • Send the code that ran, its arguments, SQL and column lineage from every plugin.

    Release 0.2.0 on GitHub

  5. 0.1.7

    Retries sign in as each tool expects.

      Release 0.1.7 on GitHub

    • 0.1.6

      Prefect tasks report the runs they launch.

        Release 0.1.6 on GitHub

      • 0.1.5

      • 0.1.4

      • 0.1.3

      • 0.1.1

        • Core client: count nesting the way the tools nest, read properties, cap batches by size.
        • Airflow: leave the session behind, and name the run a task event belongs to.
        • Dagster: forward what the run points at, not only what the context holds.
        • Prefect: name the flow and the flow run on a task event.
        • Great Expectations: forward the validation result, not the runtime that owns the rows.
        • Spark: forward the run rather than every task, and name the application on every event.
        • Land the serializer, client and five plugin fixes on main.
        • Airflow: name the task group rather than walking its copy of the DAG.

        Release 0.1.1 on GitHub

      • 0.1.0