nniback / core /admin /sandbox_pool_admin_api.py
Kyrlos11's picture
Upload 842 files
22087d0 verified
Raw History Blame Contribute Delete
9.34 kB
from fastapi import APIRouter, HTTPException, Depends, Query
from typing import Optional
from pydantic import BaseModel
from core.auth import require_admin
from core.utils.logger import logger
router = APIRouter(prefix="/admin/sandbox-pool", tags=["admin", "sandbox-pool"])
class ForceCreateRequest(BaseModel):
count: int = 1
class RemoveFromPoolRequest(BaseModel):
sandbox_ids: list[str]
delete_sandbox: bool = False
@router.get("/stats")
async def get_sandbox_pool_stats(admin: dict = Depends(require_admin)):
try:
from core.sandbox.pool_service import get_pool_service
from core.sandbox.pool_config import get_pool_config
from core.sandbox import pool_repo
service = get_pool_service()
config = get_pool_config()
pool_size = await pool_repo.get_pool_size()
stats = service.get_stats()
stats['pool_size'] = pool_size
stats['config'] = {
'enabled': config.enabled,
'min_size': config.min_size,
'max_size': config.max_size,
'replenish_threshold': config.replenish_threshold,
'check_interval': config.check_interval,
'max_age': config.max_age,
}
return stats
except Exception as e:
logger.error(f"Error getting sandbox pool stats: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.get("/health")
async def get_pool_health(admin: dict = Depends(require_admin)):
try:
from core.sandbox.pool_background import is_pool_service_running
from core.sandbox.pool_config import get_pool_config
from core.sandbox import pool_repo
config = get_pool_config()
pool_size = await pool_repo.get_pool_size()
service_running = is_pool_service_running()
health_status = "healthy"
issues = []
if not config.enabled:
health_status = "disabled"
issues.append("Pool service is disabled via configuration")
elif not service_running:
health_status = "critical"
issues.append("Background service is not running")
elif pool_size == 0:
health_status = "warning"
issues.append("Pool is empty")
elif pool_size < config.replenish_below:
health_status = "warning"
issues.append(f"Pool size ({pool_size}) below replenish threshold ({config.replenish_below})")
return {
"status": health_status,
"service_running": service_running,
"pool_enabled": config.enabled,
"pool_size": pool_size,
"min_size": config.min_size,
"replenish_threshold": config.replenish_below,
"issues": issues,
}
except Exception as e:
logger.error(f"Error getting pool health: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.get("/list")
async def list_pooled_sandboxes(
limit: int = Query(50, ge=1, le=100),
admin: dict = Depends(require_admin)
):
try:
from core.sandbox import pool_repo
sandboxes = await pool_repo.get_pooled_sandboxes(limit=limit)
return {
"count": len(sandboxes),
"sandboxes": [
{
"id": s["id"],
"external_id": s["external_id"],
"pooled_at": s.get("pooled_at"),
"created_at": s.get("created_at"),
}
for s in sandboxes
]
}
except Exception as e:
logger.error(f"Error listing pooled sandboxes: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/replenish")
async def trigger_pool_replenish(admin: dict = Depends(require_admin)):
try:
from core.sandbox.pool_service import get_pool_service
from core.sandbox import pool_repo
service = get_pool_service()
size_before = await pool_repo.get_pool_size()
created = await service.ensure_pool_size()
size_after = await pool_repo.get_pool_size()
return {
"success": True,
"sandboxes_created": created,
"pool_size_before": size_before,
"pool_size_after": size_after,
}
except Exception as e:
logger.error(f"Error triggering pool replenish: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/force-create")
async def force_create_sandboxes(
request_body: ForceCreateRequest,
admin: dict = Depends(require_admin)
):
try:
from core.sandbox.pool_service import get_pool_service
from core.sandbox import pool_repo
import asyncio
count = min(request_body.count, 10)
service = get_pool_service()
size_before = await pool_repo.get_pool_size()
tasks = [service.create_pooled_sandbox() for _ in range(count)]
results = await asyncio.gather(*tasks, return_exceptions=True)
created = []
failed = []
for i, r in enumerate(results):
if isinstance(r, Exception):
failed.append(str(r))
elif r is not None:
created.append(r)
else:
failed.append(f"Sandbox {i+1} returned None")
size_after = await pool_repo.get_pool_size()
return {
"success": True,
"requested": count,
"created_count": len(created),
"created_ids": created,
"failed_count": len(failed),
"failed_errors": failed[:5],
"pool_size_before": size_before,
"pool_size_after": size_after,
}
except Exception as e:
logger.error(f"Error force creating sandboxes: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/cleanup")
async def trigger_pool_cleanup(admin: dict = Depends(require_admin)):
try:
from core.sandbox.pool_service import get_pool_service
from core.sandbox import pool_repo
service = get_pool_service()
size_before = await pool_repo.get_pool_size()
cleaned = await service.cleanup_stale_sandboxes()
size_after = await pool_repo.get_pool_size()
return {
"success": True,
"cleaned_count": cleaned,
"pool_size_before": size_before,
"pool_size_after": size_after,
}
except Exception as e:
logger.error(f"Error triggering pool cleanup: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/remove")
async def remove_from_pool(
request_body: RemoveFromPoolRequest,
admin: dict = Depends(require_admin)
):
try:
from core.sandbox import pool_repo
from core.sandbox.sandbox import delete_sandbox
removed = []
failed = []
for sandbox_id in request_body.sandbox_ids[:20]:
try:
resource = await pool_repo.get_sandbox_by_external_id(sandbox_id)
if not resource:
failed.append({"id": sandbox_id, "error": "Not found"})
continue
if resource.get("status") != "pooled":
failed.append({"id": sandbox_id, "error": f"Status is {resource.get('status')}, not pooled"})
continue
if request_body.delete_sandbox:
try:
await delete_sandbox(sandbox_id)
except Exception as del_err:
logger.warning(f"Failed to delete sandbox {sandbox_id}: {del_err}")
await pool_repo.mark_sandbox_deleted(resource["id"])
removed.append(sandbox_id)
except Exception as e:
failed.append({"id": sandbox_id, "error": str(e)})
return {
"success": True,
"removed_count": len(removed),
"removed_ids": removed,
"failed_count": len(failed),
"failed": failed,
}
except Exception as e:
logger.error(f"Error removing sandboxes from pool: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/restart-service")
async def restart_pool_service(admin: dict = Depends(require_admin)):
try:
from core.sandbox.pool_background import stop_pool_service, start_pool_service, is_pool_service_running
was_running = is_pool_service_running()
await stop_pool_service()
await start_pool_service()
is_running = is_pool_service_running()
return {
"success": True,
"was_running": was_running,
"is_running": is_running,
"message": "Pool service restarted" if is_running else "Failed to restart pool service"
}
except Exception as e:
logger.error(f"Error restarting pool service: {e}")
raise HTTPException(status_code=500, detail=str(e))