feat: add list_applications tool for YARN application enumeration

Add a new MCP tool that queries YARN's /ws/v1/cluster/apps endpoint
through a named Connection, returning a list of ApplicationSummary
records. Bypasses the local JobStore — useful for enumerating apps
that were not submitted through this service.

API:
  list_applications(
    connection_name: str,           # required, which YARN cluster
    state: str | None = None,       # YARN state filter: NEW/NEW_SAVING/
                                    # SUBMITTED/ACCEPTED/RUNNING/
                                    # FINISHED/FAILED/KILLED
    queue: str | None = None,       # YARN queue filter
    limit: int = 100,               # cap on returned apps (YARN has no
                                    # offset-based pagination; combine
                                    # state/queue filters for big clusters)
  ) -> list[ApplicationSummary]

Implementation:
  - yarn_client.list_applications(config, *, state, queue, limit) -> list[dict]
    Returns raw YARN app dicts; raises YarnError on 4xx/5xx; returns
    [] on 404 (no apps match). Uses the existing _request helper,
    which now accepts a "params" kwarg for query strings (one-line
    additive change).
  - external_jobs.list_applications(connection_name, state, queue, limit)
    -> list[ApplicationSummary]. Looks up the Connection, builds the
    YarnClientConfig, calls the yarn_client function, maps each raw
    YARN dict to ApplicationSummary (mirroring the manual field-mapping
    style of get_job_result). The yarn_client function is imported
    as "list_applications_yarn" to avoid name collision.
  - ApplicationSummary: 12-field Pydantic model with snake_case names
    (application_id, name, user, queue, state, final_status,
    application_type, application_tags, started_time, finished_time,
    tracking_url, progress). Unused YARN fields (memorySeconds,
    vcoreSeconds, preemptedResource*, etc.) are not exposed.
  - ListApplicationsRequest: Pydantic body model with Field(description=)
    for LLM-facing schema.
  - /list_applications route registered with operation_id=
    "list_applications", placed next to the other external YARN tools.

Tests:
  - 8 new unit tests in test_external_jobs.py (happy path, state/queue/
    limit pass-through, default limit, empty list, missing connection,
    full field mapping).
  - test_mcp_routes.py: assert 23 tool routes.
  - README: list_applications row added to the Spark Executor table.

Tests: 390 passed (was 382, +8 net).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Claude
2026-07-09 13:40:20 +08:00
co-authored by Claude Fable 5
parent a5b9539663
commit 7fbad97a87
8 changed files with 363 additions and 7 deletions
+55 -3
View File
@@ -121,13 +121,14 @@ def _base_url(yarn_rm_url: str | None) -> str:
def _request(method: str, url: str, *, json_body: dict | None = None,
timeout: float = 30.0, verify: bool | str = True,
auth: httpx.Auth | None = None) -> httpx.Response:
params: dict[str, str] | None = None, timeout: float = 30.0,
verify: bool | str = True, auth: httpx.Auth | None = None) -> httpx.Response:
headers = {"Accept": "application/json"}
logger.debug(f"YARN {method} {url}" + (f" body={json_body}" if json_body else ""))
try:
resp = httpx.request(
method, url, json=json_body, headers=headers, timeout=timeout, verify=verify, auth=auth
method, url, json=json_body, params=params, headers=headers,
timeout=timeout, verify=verify, auth=auth
)
except httpx.HTTPError as exc:
logger.error(f"YARN {method} {url} failed: {exc}")
@@ -257,3 +258,54 @@ def kill_application(application_id: str, config: YarnClientConfig) -> None:
logger.error(f"YARN PUT {url} -> {resp.status_code}: {resp.text[:500]}")
raise YarnError(f"YARN kill returned HTTP {resp.status_code}: {resp.text}")
logger.info(f"YARN kill {application_id} -> ok")
def list_applications(
config: YarnClientConfig,
*,
state: str | None = None,
queue: str | None = None,
limit: int | None = None,
) -> list[dict]:
"""List YARN applications, optionally filtered.
YARN endpoint: GET /ws/v1/cluster/apps?state=...&queue=...&limit=...
Filters:
- state: YARN application state. Common values:
"NEW", "NEW_SAVING", "SUBMITTED", "ACCEPTED", "RUNNING",
"FINISHED", "FAILED", "KILLED".
Note: "FINISHED" is the umbrella state covering SUCCEEDED/FAILED/KILLED.
- queue: YARN queue name
- limit: cap on number of returned apps (YARN has no pagination;
callers that need a full enumeration should make multiple
calls with state=... filters or accept the cap)
Returns a list of YARN app dicts (each with id, name, user, queue,
state, finalStatus, applicationType, startedTime, finishedTime,
trackingUrl, progress, etc). Empty list if no apps match.
Raises YarnError on transport / 4xx / 5xx.
"""
params: dict[str, str] = {}
if state is not None:
params["state"] = state
if queue is not None:
params["queue"] = queue
if limit is not None:
params["limit"] = str(limit)
url = f"{_base_url(config.yarn_rm_url)}/ws/v1/cluster/apps"
resp = _request("GET", url, params=params,
verify=config.verify_for_httpx(),
auth=config.auth_for_httpx())
if resp.status_code == 404:
# No apps match (or RM doesn't support the endpoint)
return []
if resp.status_code >= 400:
raise YarnError(
f"YARN list applications failed: {resp.status_code} {resp.text[:200]}"
)
data = resp.json()
apps_container = data.get("apps") or {}
return apps_container.get("app", []) or []