CHRN

Autoscaling churn

Dynamic allocation adds and removes executors again and again.

Outcome: Cost
Scope: Judged on the whole run
Usual fix: A Spark configuration change
CI gate: No budget flag: gate it with --fail-on-introduced against a baseline

An example finding

Where you see it

Dashboard

The CHRN tag marks it in the action region of the dashboard, among the widgets that rank a fix. How to read it in the findings guide.

Agent (MCP)

An assistant can read this entry with the list_detectors tool and its background with get_finding_documentation. The call and the response below come from a real run of the MCP server.

Call

{
  "tool": "list_detectors",
  "arguments": {
    "type": "autoscalingChurn"
  }
}

Response

{
  "count": 1,
  "detectors": [
    {
      "type": "autoscalingChurn",
      "name": "autoscaling churn",
      "tag": "CHRN",
      "tagSharedWith": [],
      "family": "cost",
      "pitch": "Dynamic allocation adds and removes executors again and again.",
      "region": "action",
      "scope": "app",
      "fixEffort": "config",
      "thresholdSummary": "over 30% of removed executors living under 2 minutes",
      "detectors": [
        {
          "detector": "autoscalingChurn",
          "version": 2,
          "thresholds": {
            "shortLivedMs": 120000,
            "warningPct": 0.3,
            "criticalPct": 0.6,
            "minExecutors": 5
          }
        }
      ],
      "pages": {
        "catalog": "docs/detectors/autoscalingChurn.html",
        "guide": "docs/user-guide/understanding-findings.html#chrn"
      },
      "tuning": {
        "kind": "section",
        "anchor": "#bottleneck-autoscaling-churn",
        "path": "docs/tuning-reference/cluster-config.html#bottleneck-autoscaling-churn"
      },
      "gate": null
    }
  ]
}

get_finding_documentation call

{
  "tool": "get_finding_documentation",
  "arguments": {
    "type": "autoscalingChurn"
  }
}

Its response, with each Markdown body replaced by its length

{
  "type": "autoscalingChurn",
  "name": "Autoscaling Churn",
  "detectionDoc": {
    "tag": "CHRN",
    "title": "Autoscaling churn",
    "content": "<556 characters of Markdown>"
  },
  "tuningDoc": {
    "anchor": "#bottleneck-autoscaling-churn",
    "title": "Cluster Tuning",
    "content": "<15933 characters of Markdown>"
  }
}

CI (CLI)

List this entry from the command line. The output below is a real run of the command. Budgets and exit codes are in the CI guide.

sparkforensics-analyze detectors --type autoscalingChurn --format md

# Detectors (1)

## Cost

| Tag | Type | Scope | Fix | Gate | What it finds |
| --- | --- | --- | --- | --- | --- |
| CHRN | `autoscalingChurn` | app | config |  | Dynamic allocation adds and removes executors again and again. |

Report only this finding type for a run with --type:

sparkforensics-analyze <event-log> --type autoscalingChurn --impact critical,warning,info

It has no budget flag. Fail a build on a new one by comparing against a baseline with --fail-on-introduced.

Airflow

The operator has no budget parameter for it. Use fail_on_introduced with a baseline.

check_spark_job = SparkForensicsOperator(
    task_id="check_spark_job",
    log_source=...,
    backend=SubprocessAnalyzeHook(),
    baseline_log_source=...,
    fail_on_introduced="critical",
)

The operator works per run, and its summary counts findings per impact band. See the operator for its parameters.

How it is measured

Thresholds

Over 30% of removed executors living under 2 minutes.

autoscalingChurn version 2

Threshold	Default
`shortLivedMs`	120000
`warningPct`	0.3
`criticalPct`	0.6
`minExecutors`	5

Each rule's impact bands and fixed fallbacks are in the threshold tables.

Override a default per run with --thresholds; see Tuning detector thresholds.

Docs

This finding is covered in a section of a tuning page: Spark tuning reference.
Findings guide: CHRN in Understanding findings.