async def __call__(self, scope, receive, send):
http_body_bytes = await self.receive_http_body(scope, receive, send)
headers = {k.decode(): v.decode() for k, v in scope["headers"]}
# scope["router"] and scope["endpoint"] contain references to a router
# and endpoint object, respectively, which each in turn contain a
# reference to the Serve client, which cannot be serialized.
# The solution is to delete these from scope, as they will not be used.
del scope["router"]
del scope["endpoint"]
# Modify the path and root path so that reverse lookups and redirection
# work as expected. We do this here instead of in replicas so it can be
# changed without restarting the replicas.
scope["path"] = scope["path"].replace(self.path_prefix, "", 1)
scope["root_path"] = self.path_prefix
handle = self.handle.options(
method_name=headers.get("X-SERVE-CALL-METHOD".lower(),
DEFAULT.VALUE),
shard_key=headers.get("X-SERVE-SHARD-KEY".lower(), DEFAULT.VALUE),
http_method=scope["method"].upper(),
http_headers=headers)
# NOTE(edoakes): it's important that we defer building the starlette
# request until it reaches the backend replica to avoid unnecessary
# serialization cost, so we use a simple dataclass here.
request = HTTPRequestWrapper(scope, http_body_bytes)
retries = 0
backoff_time_s = 0.05
while retries < MAX_REPLICA_FAILURE_RETRIES:
object_ref = await handle.remote(request)
try:
result = await object_ref
break
except RayActorError:
logger.warning(
"Request failed due to replica failure. There are "
f"{MAX_REPLICA_FAILURE_RETRIES - retries} retries "
"remaining.")
await asyncio.sleep(backoff_time_s)
backoff_time_s *= 2
retries += 1
if isinstance(result, RayTaskError):
error_message = "Task Error. Traceback: {}.".format(result)
await Response(error_message,
status_code=500).send(scope, receive, send)
elif isinstance(result, starlette.responses.Response):
await result(scope, receive, send)
else:
await Response(result).send(scope, receive, send)
async def _send_request_to_handle(handle, scope, receive, send):
http_body_bytes = await receive_http_body(scope, receive, send)
headers = {k.decode(): v.decode() for k, v in scope["headers"]}
handle = handle.options(
method_name=headers.get("X-SERVE-CALL-METHOD".lower(), DEFAULT.VALUE),
shard_key=headers.get("X-SERVE-SHARD-KEY".lower(), DEFAULT.VALUE),
http_method=scope["method"].upper(),
http_headers=headers,
)
# scope["router"] and scope["endpoint"] contain references to a router
# and endpoint object, respectively, which each in turn contain a
# reference to the Serve client, which cannot be serialized.
# The solution is to delete these from scope, as they will not be used.
# TODO(edoakes): this can be removed once we deprecate the old API.
if "router" in scope:
del scope["router"]
if "endpoint" in scope:
del scope["endpoint"]
# NOTE(edoakes): it's important that we defer building the starlette
# request until it reaches the backend replica to avoid unnecessary
# serialization cost, so we use a simple dataclass here.
request = HTTPRequestWrapper(scope, http_body_bytes)
# Perform a pickle here to improve latency. Stdlib pickle for simple
# dataclasses are 10-100x faster than cloudpickle.
request = pickle.dumps(request)
retries = 0
backoff_time_s = 0.05
while retries < MAX_REPLICA_FAILURE_RETRIES:
object_ref = await handle.remote(request)
try:
result = await object_ref
break
except RayActorError:
logger.warning("Request failed due to replica failure. There are "
f"{MAX_REPLICA_FAILURE_RETRIES - retries} retries "
"remaining.")
await asyncio.sleep(backoff_time_s)
backoff_time_s *= 2
retries += 1
if isinstance(result, RayTaskError):
error_message = "Task Error. Traceback: {}.".format(result)
await Response(error_message,
status_code=500).send(scope, receive, send)
elif isinstance(result, starlette.responses.Response):
await result(scope, receive, send)
else:
await Response(result).send(scope, receive, send)
async def _send_request_to_handle(handle, scope, receive, send) -> str:
http_body_bytes = await receive_http_body(scope, receive, send)
# NOTE(edoakes): it's important that we defer building the starlette
# request until it reaches the replica to avoid unnecessary
# serialization cost, so we use a simple dataclass here.
request = HTTPRequestWrapper(scope, http_body_bytes)
# Perform a pickle here to improve latency. Stdlib pickle for simple
# dataclasses are 10-100x faster than cloudpickle.
request = pickle.dumps(request)
retries = 0
backoff_time_s = 0.05
loop = asyncio.get_event_loop()
# We have received all the http request conent. The next `receive`
# call might never arrive; if it does, it can only be `http.disconnect`.
client_disconnection_task = loop.create_task(receive())
while retries < MAX_REPLICA_FAILURE_RETRIES:
assignment_task = loop.create_task(handle.remote(request))
done, _ = await asyncio.wait(
[assignment_task, client_disconnection_task],
return_when=FIRST_COMPLETED)
if client_disconnection_task in done:
message = await client_disconnection_task
assert message["type"] == "http.disconnect", (
"Received additional request payload that's not disconnect. "
"This is an invalid HTTP state.")
logger.warning(
f"Client from {scope['client']} disconnected, cancelling the "
"request.")
# This will make the .result() to raise cancelled error.
assignment_task.cancel()
try:
object_ref = await assignment_task
result = await object_ref
client_disconnection_task.cancel()
break
except asyncio.CancelledError:
# Here because the client disconnected, we will return a custom
# error code for metric tracking.
return DISCONNECT_ERROR_CODE
except RayTaskError as error:
error_message = "Task Error. Traceback: {}.".format(error)
await Response(error_message,
status_code=500).send(scope, receive, send)
return "500"
except RayActorError:
logger.debug("Request failed due to replica failure. There are "
f"{MAX_REPLICA_FAILURE_RETRIES - retries} retries "
"remaining.")
await asyncio.sleep(backoff_time_s)
# Be careful about the expotential backoff scaling here.
# Assuming 10 retries, 1.5x scaling means the last retry is 38x the
# initial backoff time, while 2x scaling means 512x the initial.
backoff_time_s *= 1.5
retries += 1
else:
error_message = "Task failed with " f"{MAX_REPLICA_FAILURE_RETRIES} retries."
await Response(error_message,
status_code=500).send(scope, receive, send)
return "500"
if isinstance(result, (starlette.responses.Response, RawASGIResponse)):
await result(scope, receive, send)
return str(result.status_code)
else:
await Response(result).send(scope, receive, send)
return "200"
async def _send_request_to_handle(handle, scope, receive, send) -> str:
http_body_bytes = await receive_http_body(scope, receive, send)
headers = {k.decode(): v.decode() for k, v in scope["headers"]}
handle = handle.options(
method_name=headers.get("X-SERVE-CALL-METHOD".lower(), DEFAULT.VALUE),
shard_key=headers.get("X-SERVE-SHARD-KEY".lower(), DEFAULT.VALUE),
http_method=scope["method"].upper(),
http_headers=headers,
)
# scope["router"] and scope["endpoint"] contain references to a router
# and endpoint object, respectively, which each in turn contain a
# reference to the Serve client, which cannot be serialized.
# The solution is to delete these from scope, as they will not be used.
# TODO(edoakes): this can be removed once we deprecate the old API.
if "router" in scope:
del scope["router"]
if "endpoint" in scope:
del scope["endpoint"]
# NOTE(edoakes): it's important that we defer building the starlette
# request until it reaches the replica to avoid unnecessary
# serialization cost, so we use a simple dataclass here.
request = HTTPRequestWrapper(scope, http_body_bytes)
# Perform a pickle here to improve latency. Stdlib pickle for simple
# dataclasses are 10-100x faster than cloudpickle.
request = pickle.dumps(request)
retries = 0
backoff_time_s = 0.05
loop = asyncio.get_event_loop()
# We have received all the http request conent. The next `receive`
# call might never arrive; if it does, it can only be `http.disconnect`.
client_disconnection_task = loop.create_task(receive())
while retries < MAX_REPLICA_FAILURE_RETRIES:
assignment_task = loop.create_task(handle.remote(request))
done, _ = await asyncio.wait(
[assignment_task, client_disconnection_task],
return_when=FIRST_COMPLETED)
if client_disconnection_task in done:
message = await client_disconnection_task
assert message["type"] == "http.disconnect", (
"Received additional request payload that's not disconnect. "
"This is an invalid HTTP state.")
logger.warning(
f"Client from {scope['client']} disconnected, cancelling the "
"request.")
# This will make the .result() to raise cancelled error.
assignment_task.cancel()
try:
object_ref = await assignment_task
result = await object_ref
client_disconnection_task.cancel()
break
except asyncio.CancelledError:
# Here because the client disconnected, we will return a custom
# error code for metric tracking.
return DISCONNECT_ERROR_CODE
except RayTaskError as error:
error_message = "Task Error. Traceback: {}.".format(error)
await Response(error_message,
status_code=500).send(scope, receive, send)
return "500"
except RayActorError:
logger.warning("Request failed due to replica failure. There are "
f"{MAX_REPLICA_FAILURE_RETRIES - retries} retries "
"remaining.")
await asyncio.sleep(backoff_time_s)
# Be careful about the expotential backoff scaling here.
# Assuming 10 retries, 1.5x scaling means the last retry is 38x the
# initial backoff time, while 2x scaling means 512x the initial.
backoff_time_s *= 1.5
retries += 1
else:
error_message = ("Task failed with "
f"{MAX_REPLICA_FAILURE_RETRIES} retries.")
await Response(error_message,
status_code=500).send(scope, receive, send)
return "500"
if isinstance(result, (starlette.responses.Response, RawASGIResponse)):
await result(scope, receive, send)
return str(result.status_code)
else:
await Response(result).send(scope, receive, send)
return "200"
async def __call__(self, scope, receive, send):
"""Implements the ASGI protocol.
See details at:
https://asgi.readthedocs.io/en/latest/specs/index.html.
"""
assert scope["type"] == "http"
self.request_counter.inc(tags={"route": scope["path"]})
if scope["path"] == "/-/routes":
return await starlette.responses.JSONResponse(
self.router.route_info)(scope, receive, send)
route_prefix, handle = self.router.match_route(scope["path"],
scope["method"])
if route_prefix is None:
return await self._not_found(scope, receive, send)
http_body_bytes = await receive_http_body(scope, receive, send)
# Modify the path and root path so that reverse lookups and redirection
# work as expected. We do this here instead of in replicas so it can be
# changed without restarting the replicas.
if route_prefix != "/":
assert not route_prefix.endswith("/")
scope["path"] = scope["path"].replace(route_prefix, "", 1)
scope["root_path"] = route_prefix
headers = {k.decode(): v.decode() for k, v in scope["headers"]}
handle = handle.options(
method_name=headers.get("X-SERVE-CALL-METHOD".lower(),
DEFAULT.VALUE),
shard_key=headers.get("X-SERVE-SHARD-KEY".lower(), DEFAULT.VALUE),
http_method=scope["method"].upper(),
http_headers=headers)
# NOTE(edoakes): it's important that we defer building the starlette
# request until it reaches the backend replica to avoid unnecessary
# serialization cost, so we use a simple dataclass here.
request = HTTPRequestWrapper(scope, http_body_bytes)
retries = 0
backoff_time_s = 0.05
while retries < MAX_REPLICA_FAILURE_RETRIES:
object_ref = await handle.remote(request)
try:
result = await object_ref
break
except RayActorError:
logger.warning(
"Request failed due to replica failure. There are "
f"{MAX_REPLICA_FAILURE_RETRIES - retries} retries "
"remaining.")
await asyncio.sleep(backoff_time_s)
backoff_time_s *= 2
retries += 1
if isinstance(result, RayTaskError):
error_message = "Task Error. Traceback: {}.".format(result)
await Response(
error_message, status_code=500).send(scope, receive, send)
elif isinstance(result, starlette.responses.Response):
await result(scope, receive, send)
else:
await Response(result).send(scope, receive, send)
