AmcRedfishDiagPlugin

[alexandraBara]

Summary

-AmcRedfishDiagPlugin collects AMC Redfish diagnostic archives by SSH’ing to the BMC and curling the AMC Redfish service
-It discovers a LogService that advertises CollectDiagnosticData, posts the job, polls the task, and downloads the attachment.
-Collects OEM AllLogs tar + Managers dump

Test plan

• pytest test/unit
• pytest test/functional (if applicable)
• pre-commit run --all-files

Checklist

• Added/updated tests (or explained why not)
• Updated docs/README if behavior changed
• No secrets or credentials committed

Sample connection_config.json:

{
  "RedfishSshProxyConnectionManager": {
    "host": "192.168.31.1",
    "port": 80,
    "use_https": false,
    "timeout_seconds": 60,
    "api_root": "redfish/v1",
    "ssh": {
      "hostname": "<BMC_hostname/IP>",
      "username": "...",
      "password": "...",
      "port": 22
    }
  }
}

Sample amc_redfish_diag_plugin_config.json:

{
  "name": "AmcRedfishDiagnostics",
  "desc": "Collect AMC Manager and System diagnostic bundles through BMC SSH proxy",
  "global_args": {},
  "plugins": {
    "AmcRedfishDiagPlugin": {
      "collection": true,
      "analysis": true,
      "collection_args": {
        "manager_ids": [],
        "system_ids": [],
        "collections": [
          {
            "root": "Managers",
            "diagnostic_data_type": "Manager"
          },
          {
            "root": "Systems",
            "diagnostic_data_type": "OEM",
            "oem_data_type": "AllLogs"
          }
        ],
        "task_timeout_s": 1800
      },
      "analysis_args": {
        "require_all_success": true
      }
    }
  },
  "result_collators": {}
}

Run:

node-scraper  --connection-config ~/configs/connection_config.json   --plugin-config  amc_redfish_diag_plugin_config.json

[amd-ansonyim]

one thing but other that LGTM

[amd-ansonyim]

This sets OK as long as any job succeeds. If the Manager dump succeeds but the Systems AllLogs collection fails, the top-level result is OK even though half the data is missing. Users relying on the status to gate downstream processing will be misled. A stricter check — e.g., OK only when success_count == len(results) and PARTIAL/WARNING otherwise — would be safer.

[alexandraBara]

good point, ill take a look