Cloud Saas example deployment wrong

[111andre111]

Overview

Expected Behavior

Deployment should be created without error.

Current Behavior

The documented way of creating a new deployment may work for ECE, but it doesn't work with Cloud SaaS.
I found at least 2 references, maybe there exist more in the documentation:
https://www.elastic.co/guide/en/ecctl/current/ecctl_deployment_create.html
https://github.com/elastic/ecctl/blob/master/docs/ecctl_deployment_create.md#examples

ecctl deployment create --file ./deployment_example.json --name=my-new-deployment --version=7.4.1
It results in error message:
Instance configuration [data.default] does not exist or is not available to you. The instance configuration must be one of the available configurations. Use the Instance Configuration API to get a list of available configurations.

Possible Solution

I did lots of investigation and came up with this solution creating a deployment with elasticsearch, kibana and apm:

cat deployment_example_gcp.json

{
  "cluster_name": "",
  "plan": {
    "cluster_topology": [
      {
        "node_type": {
          "data": true,
          "master": true,
          "ingest": true
        },
        "instance_configuration_id": "gcp.data.highio.1",
        "zone_count": 3,
        "size": {
          "resource": "memory",
          "value": 65536
        },
        "elasticsearch": {
          "enabled_built_in_plugins": []
        }
      },
      {
        "node_type": {
          "data": false,
          "master": false,
          "ingest": false,
          "ml": true
        },
        "instance_configuration_id": "gcp.ml.1",
        "zone_count": 1,
        "size": {
          "resource": "memory",
          "value": 16384
        },
        "elasticsearch": {
          "enabled_built_in_plugins": []
        }
      }
    ],
    "elasticsearch": {
      "version": "7.4.0"
    },
    "transient": {},
    "deployment_template": {
      "id": "gcp-io-optimized"
    }
  },
  "settings": {
    "dedicated_masters_threshold": 6
  },
  "kibana": {
    "plan": {
      "cluster_topology": [
        {
          "instance_configuration_id": "gcp.kibana.1",
          "zone_count": 1,
          "size": {
            "resource": "memory",
            "value": 1024
          }
        }
      ],
      "kibana": {
        "version": "7.4.0"
      }
    }
  }
}


ecctl deployment create --file ./deployment_example_gcp.json --name=my-new-deployment --version=7.4.1

Context

I found at least another issue:

At the moment the plan path plan.cluster_topology.instance_configuration_id is not well documented for Cloud SaaS implementation.

I did not find a command which can list all available possibilities.

I found at least these ones:

    aws-appsearch-dedicated
    aws-compute-optimized
    aws-cross-cluster-search
    aws-hot-warm
    aws-io-optimized
    aws-memory-optimized
    default
    gcp-appsearch-dedicated
    gcp-compute-optimized
    gcp-cross-cluster-search
    gcp-hot-warm
    gcp-io-optimized
    gcp-memory-optimized
    azure-appsearch-dedicated
    azure-compute-optimized
    azure-cross-cluster-search
    azure-hot-warm
    azure-io-optimized
    azure-memory-optimized

Additionally I tried to monitor this deployment plan, but it didn't seem to work as well:
ecctl deployment elasticsearch plan monitor <cloudID>

Reference:
#129

[karencfv]

Hi @111andre111! First of all thank you so much for taking the time to try out ecctl and give us feedback 🙇‍♀

About the ecctl deployment create command, I'm currently working on a related issue -> #116 and will take the feedback here into account and implement it in that PR since I'll be changing the help text anyway.

I think the command you're looking for regarding the deployment templates is ecctl platform deployment-template list . Maybe it'd be a good idea to reference that command within the ecctl deployment create help text 🤔

With the ecctl deployment elasticsearch plan monitor <cloudID> command, what were the issues you ran into?

[marclop]

I think we should add a reference on how to find the instance configuration IDs and deployment templates, it can get confusing otherwise.

Also, by specifying the --track flag on ecctl deployment create will automatically track / monitor the creation of the deployment and all of its downstream resources.