Duplicated plugin name in json status

[johnSchnake]

What steps did you take and what happened:
sonobuoy status --json | jq

Should have a list of the plugins and their aggregate data. However, you can see in the output below that it duplicated the plugin and only put the data really into one of them.

{
  "plugins": [
    {
      "plugin": "daemonset-worker",
      "node": "ip-172-31-30-196.us-west-2.compute.internal",
      "status": "complete",
      "result-status": "failed",
      "result-counts": {
        "failed": 12,
        "passed": 32,
        "skipped": 4
      }
    },
    {
      "plugin": "daemonset-worker",
      "node": "ip-172-31-47-230.us-west-2.compute.internal",
      "status": "complete",
      "result-status": "",
      "result-counts": null
    }
  ],

It also didnt list the daemonset-master info (which didnt really run since none of the nodes were schedulable but I would think should be in that list).

FWIW the plugin daemonset-worker was targeting nodes without the master label on them; daemonset-master the opposite.

Sonobuoy v0.16.4

[johnSchnake]

I just realized that the issue is that the node is being set whereas the plugin itself isn't node specific. So I guess the plugin-node logic is adding 2 values (with no result-counts) and then the postprocessing logic is probably just finding one and adding the count data. Something along those lines.

Will have to check what the original intent was here; I'm not certain whether or not the intent was to have the values in that list being dependent on the node at all. I thought it was a rollup of that plugin across the cluster. TBD

[johnSchnake]

Easiest way to see this is to set up a 2 node kind cluster and run

sonobuoy run --plugin systemd-logs --wait
sonobuoy status --json | jq