Catalog Worker

The Catalog Worker runs under the aegis of Red Hat Connect (rhc). It can be run in a vm or container and talks to an on-prem Ansible Tower. It can

• Collect inventory objects from Ansible Tower
• Launch and monitor jobs on Ansible Tower.

The Red Hat Connect (rhc) subscribes to a specific topic on the Cloud Connector based on its unqiue guid. When a task needs to be done on the client the Cloud Connector sends a small message packet to the RHC which transfers the message to the Catalog Worker via GRPC. The message includes the url to get the task details, a date time stamp and the kind of task.

{
    "url": "http://cloud.redhat.com/api/catalog-inventory/v3.0/tasks/xxxx",
    "kind": "catalog",
    "sent": "2020-10-03T12:34:56Z"
} 

Once the catalog worker gets this message it looks at the URL and fetches the task details.

The client updates the task after it has finished processing. The client can either send the response directly to the task#result or it can upload a compress tar file to the upload service. Since the inventory data tends to be big we usually upload that via a compressed tar file. For other simple requests we directly update the task#results.

A task is a collection of jobs alongwith result format and upload url.

e.g.

{
    "response_format": "tar|json",
    "upload_url": "https://cloud.redhat.com/api/v1/ingress/upload"
    "jobs": [{
        "href_slug": "/api/v2/job_templates",
        "method": "get",
        "fetch_all_pages": true,
        "apply_filter": "results[].{id:id, inventory:inventory, type:type, url:url}"
        "fetch_related": [{
            "href_slug": "survey_spec",
            "predicate": "survey_enabled"
        }]
    }]
}

Parameters for Catalog Worker

The Parameters for the catalog worker are stored in /etc/yggdrasil/workers/catalog.conf

Task Parameters

Keyword	Description	Example
response_format	Compressed tar file or json	tar
upload_url	The URL of the upload service	https://cloud.redhat.com/api/ingress/v1/upload
jobs	An array of jobs for this task	See example below

Job Parameters

Keyword	Description	Example
href_slug	The Partial URL (required)	/api/v2/job_templates
method	One of get/post/monitor/launch (required)	get
fetch_all_pages	Fetch all pages from Tower for a URL	true
apply_filter	JMES Path filter to trim data	results[].{id:id, type:type, created:created,name:name
params	Post Params or Query Params
fetch_related	Optionally fetch other related objects

The list of inventory objects to be collected from the tower is sent from the cloud.redhat.com. The list of objects needed by catalog are

1. Job Templates
2. Inventories
3. Credentials
4. Credential Types
5. Survey Spec
6. Workflow Templates
7. Workflow Template Nodes

General Workflow