Robert Wojciechowski
6b94242110
* feat: add Runpod serverless image collector pools * fix: prevent worker settings clipping and remove preview copy * fix: initialize worker settings before mounting the form * fix: validate complete serverless requests against 10 MB limit * fix: simplify serverless cards and keep status dots circular * fix: match serverless endpoint box to master card * fix: preserve serverless capacity and recovery across job failures Retain durable reservations until provider termination or explicit recovery, protect unresolved pools and credentials, and add scoped reconciliation. Consolidate lifecycle and secret policies and split prompt/media stages with regression coverage. * fix(serverless): disable Clear saved API key while the worker is enabled The checkbox read "Clear saved API key (disable pool first)", but the panel never defines "pool" and saving with it ticked on an enabled worker was rejected by the backend. - Disable the checkbox while the worker is enabled, with a tooltip saying to turn the worker off first. Also disable it when no key is saved. - Use "worker" instead of "pool" in user-facing serverless errors, the serverless guide and the README entry. * feat(serverless): replace Concurrent jobs with a single Images per run setting - Remove the separate Concurrent jobs field; rename Jobs per run to Images per run (1-16) with a tooltip explaining that Runpod's max workers decides how many GPUs run them. - Backend uses a fixed per-worker limit of 16 outstanding jobs, so all images of a run are submitted at once. Unresolved jobs still count against it. The limit is constant, so it never has to change while jobs are unresolved. - Legacy 'concurrency' in saved configs and older browser tabs is accepted and ignored/dropped; it no longer triggers the busy guard. - Revert the enable toggle and show the reason when the backend rejects it. - Trim a pasted api.runpod.ai/v2/<id> URL to the endpoint ID; plainer endpoint-ID error otherwise. - Confirm before a type change deletes a saved Runpod API key. - Update docs and tests.
99 lines
4.4 KiB
Python
99 lines
4.4 KiB
Python
"""Master-only Runpod pool monitoring. Health checks never submit inference."""
|
|
from aiohttp import web
|
|
import server
|
|
|
|
from ..utils.config import load_config
|
|
from .orchestration.serverless_runtime import get_manager
|
|
from .orchestration.runpod_client import RunpodError
|
|
|
|
|
|
def _pool(worker_id):
|
|
return next((w for w in load_config().get("workers", [])
|
|
if w.get("id") == worker_id and w.get("type") == "serverless"), None)
|
|
|
|
|
|
async def _status(worker_id, fresh=False):
|
|
worker = _pool(worker_id)
|
|
if worker is None:
|
|
return web.json_response({"ok": False, "message": "Serverless worker not found"}, status=404)
|
|
jobs = []
|
|
try:
|
|
manager = get_manager()
|
|
jobs = manager.public_jobs(worker_id)
|
|
result = await manager.health(worker, fresh=fresh)
|
|
return web.json_response(dict(result, jobs=jobs))
|
|
except Exception as exc:
|
|
message = str(exc) if isinstance(exc, RunpodError) else "Check the saved endpoint ID and API key"
|
|
return web.json_response({"ok": False, "queueRemaining": 0, "message": message, "jobs": jobs})
|
|
|
|
|
|
@server.PromptServer.instance.routes.post("/distributed/serverless/test")
|
|
async def test_serverless_endpoint(request):
|
|
try:
|
|
data = await request.json()
|
|
worker_id = data.get("worker_id")
|
|
except Exception:
|
|
return web.json_response({"ok": False, "message": "Invalid request"}, status=400)
|
|
return await _status(worker_id, fresh=True)
|
|
|
|
|
|
@server.PromptServer.instance.routes.get("/distributed/serverless/status/{worker_id}")
|
|
async def serverless_status_endpoint(request):
|
|
return await _status(request.match_info["worker_id"])
|
|
|
|
|
|
@server.PromptServer.instance.routes.post("/distributed/serverless/cancel")
|
|
async def cancel_serverless_endpoint(request):
|
|
try:
|
|
data = await request.json()
|
|
worker_id = data.get("worker_id")
|
|
except Exception:
|
|
return web.json_response({"ok": False, "message": "Invalid request"}, status=400)
|
|
if _pool(worker_id) is None:
|
|
return web.json_response({"ok": False, "message": "Serverless worker not found"}, status=404)
|
|
try:
|
|
manager = get_manager()
|
|
count = await manager.cancel(worker_id=worker_id)
|
|
return web.json_response({"ok": True, "cancelled": count, "jobs": manager.public_jobs(worker_id),
|
|
"message": "Cancellation requested for this master's jobs only"})
|
|
except Exception:
|
|
return web.json_response({"ok": False, "message": "Cancellation could not be confirmed; check provider jobs"}, status=502)
|
|
|
|
|
|
async def _recover(request, *, acknowledge=False):
|
|
try:
|
|
data = await request.json()
|
|
worker_id = data.get("worker_id")
|
|
if not isinstance(worker_id, str) or not worker_id.strip():
|
|
raise ValueError()
|
|
record_id = data.get("record_id")
|
|
if acknowledge:
|
|
if not isinstance(record_id, str) or not record_id.strip() or data.get("provider_checked") is not True:
|
|
raise ValueError()
|
|
except Exception:
|
|
return web.json_response({"ok": False, "message": "Provide a worker ID; manual resolution also requires a record ID and provider_checked: true"}, status=400)
|
|
if _pool(worker_id) is None:
|
|
return web.json_response({"ok": False, "message": "Serverless worker not found"}, status=404)
|
|
try:
|
|
manager = get_manager()
|
|
if acknowledge:
|
|
await manager.acknowledge_recovery(worker_id, record_id, provider_checked=True)
|
|
result = {"message": "Recovery acknowledged by operator; not provider-confirmed cancellation"}
|
|
else:
|
|
result = {"resolved": await manager.reconcile(worker_id=worker_id)}
|
|
return web.json_response(dict(result, ok=True, jobs=manager.public_jobs(worker_id)))
|
|
except ValueError:
|
|
return web.json_response({"ok": False, "message": "Unresolved inactive job not found for this worker"}, status=409)
|
|
except Exception:
|
|
return web.json_response({"ok": False, "message": "Recovery could not be confirmed; retained jobs remain protected"}, status=502)
|
|
|
|
|
|
@server.PromptServer.instance.routes.post("/distributed/serverless/reconcile")
|
|
async def reconcile_serverless_endpoint(request):
|
|
return await _recover(request)
|
|
|
|
|
|
@server.PromptServer.instance.routes.post("/distributed/serverless/resolve")
|
|
async def resolve_serverless_endpoint(request):
|
|
return await _recover(request, acknowledge=True)
|