list_detectors

Reference for the list_detectors tool of the SparkForensics MCP server.

Every finding type the detectors can report, independent of any run: what it finds, how it is grouped, its default thresholds, where its docs live and the budget flag that gates it. An agent can call it to learn what SparkForensics checks before it loads a run. The same list is on the command line as sparkforensics-analyze detectors and as the detector catalog pages of this site.

Parameter, optional:

• type: one finding type value, such as "skew". An unknown type fails input validation. Without it the tool returns every type.

Example call:

{ "name": "list_detectors", "arguments": { "type": "skew" } }

Example response:

{
  "count": 1,
  "detectors": [
    {
      "type": "skew",
      "name": "task skew",
      "tag": "SKEW",
      "tagSharedWith": [],
      "family": "speed",
      "pitch": "One slow task holds a whole stage back while the rest of the cluster waits.",
      "region": "action",
      "scope": "stage",
      "fixEffort": "code",
      "thresholdSummary": "P95 task time over 3× the median (the longest task on stages under 20 tasks), and the slow tasks read correspondingly more data",
      "detectors": [
        {
          "detector": "skew",
          "version": 1,
          "thresholds": {
            "ratioWarn": 3,
            "minTasksForP95": 20,
            "dataShareMin": 0.5,
            "floorPctWarn": 0.005
          }
        }
      ],
      "pages": {
        "catalog": "docs/detectors/skew.html",
        "guide": "docs/user-guide/understanding-findings.html#skew"
      },
      "tuning": {
        "kind": "page",
        "anchor": "#bottleneck-skew",
        "path": "docs/tuning-reference/bottleneck-skew.html#bottleneck-skew"
      },
      "gate": {
        "cliFlag": "--max-skew",
        "airflowParam": "max_skew_ratio"
      }
    }
  ]
}

Each record has:

• type, name, tag and tagSharedWith, the other types that share the tag.
• family and pitch: the outcome group and one sentence on what the finding means.
• region: "reference" for the types the dashboard shows below its action widgets, "action" for the rest.
• scope and fixEffort: where the detector looks, and whether the fix is a setting (config) or a change to the job (code).
• thresholdSummary and detectors: the clean-check criterion, and one entry per detector that emits the type with its version and default thresholds. A type can have several entries (configAudit has three). A server started with --thresholds still lists the defaults here; its analyzing tools report the tuned values.
• pages: the catalog page of the type and its entry in the findings guide, as paths under the site root.
• tuning: where the tuning reference covers the type, as a page of its own, a page shared with other types, a section of another page or a general chapter. It is null when no single anchor covers the type.
• gate: the CLI budget flag and the Airflow operator parameter that gate the type directly. It is null for the types with no budget; those are gated with --fail-on-introduced against a baseline.

For the detection and tuning text of one type, call get_finding_documentation.