Bug 1597765

Description of problem:
Checking Capsule status and Services tab, I see:


Pulp node
  Version             1.3.0
Pulp server version   2.13.4.10
Database connection   OK
Messaging connection  OK
Workers               6

While I configured the Capsule with PULP_CONCURRENCY 4 and really there are 4 pulp workers running.

The reason is, katello counts also resource manager and scheduler "worker" among pulp celery workers. Since it fetches the data via API request:

curl -k https://pmoravec-caps63.gsslab.brq2.redhat.com/pulp/api/v2/status/ | python -m json.tool
..
{
    "api_version": "2",
    "database_connection": {
        "connected": true
    },
    "known_workers": [
        {
            "_id": "reserved_resource_worker-2.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:16Z"
        },
        {
            "_id": "reserved_resource_worker-0.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:16Z"
        },
        {
            "_id": "reserved_resource_worker-1.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:18Z"
        },
        {
            "_id": "reserved_resource_worker-3.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:26Z"
        },
        {
            "_id": "resource_manager.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:21Z"
        },
        {
            "_id": "scheduler.brq2.redhat.com",
            "_ns": "workers",
            "last_heartbeat": "2018-07-03T14:52:19Z"
        }
    ],
    "messaging_connection": {
        "connected": true
    },
    "versions": {
        "platform_version": "2.13.4.10"
    }
}

and prints @pulp_status['known_workers'].size . See

/opt/theforeman/tfm/root/usr/share/gems/gems/katello-3.4.5.74/app/views/smart_proxies/pulp_status.html.erb

for the code.

The number is correct from categorization point of view (in the above output, both scheduler and resource_manager are marked as workers), but it is confusing from end-user point of view who set pulp concurrency to 4 and thinks there are 4 workers. And from pulp concurrency point of view, there really are 4 workers only.


Version-Release number of selected component (if applicable):
tfm-rubygem-katello-3.4.5.74-1.el7sat.noarch


How reproducible:
100%


Steps to Reproduce:
1. Have an external Capsule, play with PULP_CONCURRENCY in /etc/default/pulp_workers (and restart pulp services afterwards) there.

2. Check Capsule service status and # of pulp workers


Actual results:
2. WebUI shows 2 more workers than PULP_CONCURRENCY.


Expected results:
2. The numbers to match.


Additional info: