Skip to content

Deployment Process

Deploying a punch first requires to deploy a kast instance with the required components. Refer to the kast documentation to understand the kast deployment logic and process.


Deploying kast is very similar to deploying a VM punch. You first define a few configuration files, and the kast deployer takes care of deploying everything to the target servers. If you are used to the punch deployer, you will be at ease with the kast deployer.

The detailed process is as follows:

  1. Select the third-party components required by your punch solution.
    For example : kafka, minio, clickhouse and grafana.

  2. Define and set up your hosting infrastructure.
    This typically consists in creating the required VMs (or native servers) with the required CPU, RAM and disk resources, making sure the networking, NTP and DNS are properly setup.

  3. Deploy these components using the Kast deployer.
    Refer to the Kast Getting Started Guide for a quick overview.

  4. Install a Punch Console.
    The punch console is a lightweight command line tool that provides you with the ops commands to manage your platform and applications. It runs on your installation server, your laptop or a bastion hop depending on your installation mode (offline or online).

  5. Install the Punch Images needed to start the punch services and applications.

  6. Install the Punch Services. This extends the Kubernetes cluster with REST apis and a punch kubernetes operator that simplifies significantly the management of punch and Kast data pipelines.

  7. Configure your platform typically with Kafka topics, Kibana or Grafana dashboards.

  8. Populate the Punch Artefact Server with Punch Resources such as parsers, enrichment files, geo tile maps etc.

  9. Start your punchlines and plans using the Punch Operator