diff options
Diffstat (limited to 'synapse/http')
| -rw-r--r-- | synapse/http/client.py | 2 | ||||
| -rw-r--r-- | synapse/http/endpoint.py | 12 | ||||
| -rw-r--r-- | synapse/http/matrixfederationclient.py | 15 | ||||
| -rw-r--r-- | synapse/http/request_metrics.py | 13 | ||||
| -rw-r--r-- | synapse/http/server.py | 87 | ||||
| -rw-r--r-- | synapse/http/servlet.py | 56 | ||||
| -rw-r--r-- | synapse/http/site.py | 211 |
7 files changed, 268 insertions, 128 deletions
diff --git a/synapse/http/client.py b/synapse/http/client.py index 3771e0b3f6..ab4fbf59b2 100644 --- a/synapse/http/client.py +++ b/synapse/http/client.py @@ -42,7 +42,7 @@ from twisted.web.http_headers import Headers from synapse.api.errors import Codes, HttpResponseException, SynapseError from synapse.http import cancelled_to_request_timed_out_error, redact_uri from synapse.http.endpoint import SpiderEndpoint -from synapse.util.async import add_timeout_to_deferred +from synapse.util.async_helpers import add_timeout_to_deferred from synapse.util.caches import CACHE_SIZE_FACTOR from synapse.util.logcontext import make_deferred_yieldable diff --git a/synapse/http/endpoint.py b/synapse/http/endpoint.py index d65daa72bb..b0c9369519 100644 --- a/synapse/http/endpoint.py +++ b/synapse/http/endpoint.py @@ -26,7 +26,6 @@ from twisted.names.error import DNSNameError, DomainError logger = logging.getLogger(__name__) - SERVER_CACHE = {} # our record of an individual server which can be tried to reach a destination. @@ -103,15 +102,16 @@ def parse_and_validate_server_name(server_name): return host, port -def matrix_federation_endpoint(reactor, destination, ssl_context_factory=None, +def matrix_federation_endpoint(reactor, destination, tls_client_options_factory=None, timeout=None): """Construct an endpoint for the given matrix destination. Args: reactor: Twisted reactor. destination (bytes): The name of the server to connect to. - ssl_context_factory (twisted.internet.ssl.ContextFactory): Factory - which generates SSL contexts to use for TLS. + tls_client_options_factory + (synapse.crypto.context_factory.ClientTLSOptionsFactory): + Factory which generates TLS options for client connections. timeout (int): connection timeout in seconds """ @@ -122,13 +122,13 @@ def matrix_federation_endpoint(reactor, destination, ssl_context_factory=None, if timeout is not None: endpoint_kw_args.update(timeout=timeout) - if ssl_context_factory is None: + if tls_client_options_factory is None: transport_endpoint = HostnameEndpoint default_port = 8008 else: def transport_endpoint(reactor, host, port, timeout): return wrapClientTLS( - ssl_context_factory, + tls_client_options_factory.get_options(host), HostnameEndpoint(reactor, host, port, timeout=timeout)) default_port = 8448 diff --git a/synapse/http/matrixfederationclient.py b/synapse/http/matrixfederationclient.py index bf1aa29502..b34bb8e31a 100644 --- a/synapse/http/matrixfederationclient.py +++ b/synapse/http/matrixfederationclient.py @@ -43,7 +43,7 @@ from synapse.api.errors import ( from synapse.http import cancelled_to_request_timed_out_error from synapse.http.endpoint import matrix_federation_endpoint from synapse.util import logcontext -from synapse.util.async import add_timeout_to_deferred +from synapse.util.async_helpers import add_timeout_to_deferred from synapse.util.logcontext import make_deferred_yieldable logger = logging.getLogger(__name__) @@ -61,14 +61,14 @@ MAX_SHORT_RETRIES = 3 class MatrixFederationEndpointFactory(object): def __init__(self, hs): - self.tls_server_context_factory = hs.tls_server_context_factory + self.tls_client_options_factory = hs.tls_client_options_factory def endpointForURI(self, uri): destination = uri.netloc return matrix_federation_endpoint( reactor, destination, timeout=10, - ssl_context_factory=self.tls_server_context_factory + tls_client_options_factory=self.tls_client_options_factory ) @@ -133,7 +133,7 @@ class MatrixFederationHttpClient(object): failures, connection failures, SSL failures.) """ if ( - self.hs.config.federation_domain_whitelist and + self.hs.config.federation_domain_whitelist is not None and destination not in self.hs.config.federation_domain_whitelist ): raise FederationDeniedError(destination) @@ -439,7 +439,7 @@ class MatrixFederationHttpClient(object): defer.returnValue(json.loads(body)) @defer.inlineCallbacks - def get_json(self, destination, path, args={}, retry_on_dns_fail=True, + def get_json(self, destination, path, args=None, retry_on_dns_fail=True, timeout=None, ignore_backoff=False): """ GETs some json from the given host homeserver and path @@ -447,7 +447,7 @@ class MatrixFederationHttpClient(object): destination (str): The remote server to send the HTTP request to. path (str): The HTTP path. - args (dict): A dictionary used to create query strings, defaults to + args (dict|None): A dictionary used to create query strings, defaults to None. timeout (int): How long to try (in ms) the destination for before giving up. None indicates no timeout and that the request will @@ -702,6 +702,9 @@ def check_content_type_is_json(headers): def encode_query_args(args): + if args is None: + return b"" + encoded_args = {} for k, vs in args.items(): if isinstance(vs, string_types): diff --git a/synapse/http/request_metrics.py b/synapse/http/request_metrics.py index 588e280571..72c2654678 100644 --- a/synapse/http/request_metrics.py +++ b/synapse/http/request_metrics.py @@ -15,6 +15,7 @@ # limitations under the License. import logging +import threading from prometheus_client.core import Counter, Histogram @@ -111,6 +112,9 @@ in_flight_requests_db_sched_duration = Counter( # The set of all in flight requests, set[RequestMetrics] _in_flight_requests = set() +# Protects the _in_flight_requests set from concurrent accesss +_in_flight_requests_lock = threading.Lock() + def _get_in_flight_counts(): """Returns a count of all in flight requests by (method, server_name) @@ -120,7 +124,8 @@ def _get_in_flight_counts(): """ # Cast to a list to prevent it changing while the Prometheus # thread is collecting metrics - reqs = list(_in_flight_requests) + with _in_flight_requests_lock: + reqs = list(_in_flight_requests) for rm in reqs: rm.update_metrics() @@ -154,10 +159,12 @@ class RequestMetrics(object): # to the "in flight" metrics. self._request_stats = self.start_context.get_resource_usage() - _in_flight_requests.add(self) + with _in_flight_requests_lock: + _in_flight_requests.add(self) def stop(self, time_sec, request): - _in_flight_requests.discard(self) + with _in_flight_requests_lock: + _in_flight_requests.discard(self) context = LoggingContext.current_context() diff --git a/synapse/http/server.py b/synapse/http/server.py index 6dacb31037..2d5c23e673 100644 --- a/synapse/http/server.py +++ b/synapse/http/server.py @@ -25,8 +25,9 @@ from canonicaljson import encode_canonical_json, encode_pretty_printed_json, jso from twisted.internet import defer from twisted.python import failure -from twisted.web import resource, server +from twisted.web import resource from twisted.web.server import NOT_DONE_YET +from twisted.web.static import NoRangeStaticProducer from twisted.web.util import redirectTo import synapse.events @@ -37,10 +38,13 @@ from synapse.api.errors import ( SynapseError, UnrecognizedRequestError, ) -from synapse.http.request_metrics import requests_counter from synapse.util.caches import intern_dict -from synapse.util.logcontext import LoggingContext, PreserveLoggingContext -from synapse.util.metrics import Measure +from synapse.util.logcontext import preserve_fn + +if PY3: + from io import BytesIO +else: + from cStringIO import StringIO as BytesIO logger = logging.getLogger(__name__) @@ -60,11 +64,10 @@ HTML_ERROR_TEMPLATE = """<!DOCTYPE html> def wrap_json_request_handler(h): """Wraps a request handler method with exception handling. - Also adds logging as per wrap_request_handler_with_logging. + Also does the wrapping with request.processing as per wrap_async_request_handler. The handler method must have a signature of "handle_foo(self, request)", - where "self" must have a "clock" attribute (and "request" must be a - SynapseRequest). + where "request" must be a SynapseRequest. The handler must return a deferred. If the deferred succeeds we assume that a response has been sent. If the deferred fails with a SynapseError we use @@ -108,24 +111,23 @@ def wrap_json_request_handler(h): pretty_print=_request_user_agent_is_curl(request), ) - return wrap_request_handler_with_logging(wrapped_request_handler) + return wrap_async_request_handler(wrapped_request_handler) def wrap_html_request_handler(h): """Wraps a request handler method with exception handling. - Also adds logging as per wrap_request_handler_with_logging. + Also does the wrapping with request.processing as per wrap_async_request_handler. The handler method must have a signature of "handle_foo(self, request)", - where "self" must have a "clock" attribute (and "request" must be a - SynapseRequest). + where "request" must be a SynapseRequest. """ def wrapped_request_handler(self, request): d = defer.maybeDeferred(h, self, request) d.addErrback(_return_html_error, request) return d - return wrap_request_handler_with_logging(wrapped_request_handler) + return wrap_async_request_handler(wrapped_request_handler) def _return_html_error(f, request): @@ -170,46 +172,26 @@ def _return_html_error(f, request): finish_request(request) -def wrap_request_handler_with_logging(h): - """Wraps a request handler to provide logging and metrics +def wrap_async_request_handler(h): + """Wraps an async request handler so that it calls request.processing. + + This helps ensure that work done by the request handler after the request is completed + is correctly recorded against the request metrics/logs. The handler method must have a signature of "handle_foo(self, request)", - where "self" must have a "clock" attribute (and "request" must be a - SynapseRequest). + where "request" must be a SynapseRequest. - As well as calling `request.processing` (which will log the response and - duration for this request), the wrapped request handler will insert the - request id into the logging context. + The handler may return a deferred, in which case the completion of the request isn't + logged until the deferred completes. """ @defer.inlineCallbacks - def wrapped_request_handler(self, request): - """ - Args: - self: - request (synapse.http.site.SynapseRequest): - """ + def wrapped_async_request_handler(self, request): + with request.processing(): + yield h(self, request) - request_id = request.get_request_id() - with LoggingContext(request_id) as request_context: - request_context.request = request_id - with Measure(self.clock, "wrapped_request_handler"): - # we start the request metrics timer here with an initial stab - # at the servlet name. For most requests that name will be - # JsonResource (or a subclass), and JsonResource._async_render - # will update it once it picks a servlet. - servlet_name = self.__class__.__name__ - with request.processing(servlet_name): - with PreserveLoggingContext(request_context): - d = defer.maybeDeferred(h, self, request) - - # record the arrival of the request *after* - # dispatching to the handler, so that the handler - # can update the servlet name in the request - # metrics - requests_counter.labels(request.method, - request.request_metrics.name).inc() - yield d - return wrapped_request_handler + # we need to preserve_fn here, because the synchronous render method won't yield for + # us (obviously) + return preserve_fn(wrapped_async_request_handler) class HttpServer(object): @@ -272,7 +254,7 @@ class JsonResource(HttpServer, resource.Resource): """ This gets called by twisted every time someone sends us a request. """ self._async_render(request) - return server.NOT_DONE_YET + return NOT_DONE_YET @wrap_json_request_handler @defer.inlineCallbacks @@ -413,8 +395,7 @@ def respond_with_json(request, code, json_object, send_cors=False, return if pretty_print: - json_bytes = (encode_pretty_printed_json(json_object) + "\n" - ).encode("utf-8") + json_bytes = encode_pretty_printed_json(json_object) + b"\n" else: if canonical_json or synapse.events.USE_FROZEN_DICTS: # canonicaljson already encodes to bytes @@ -450,8 +431,12 @@ def respond_with_json_bytes(request, code, json_bytes, send_cors=False, if send_cors: set_cors_headers(request) - request.write(json_bytes) - finish_request(request) + # todo: we can almost certainly avoid this copy and encode the json straight into + # the bytesIO, but it would involve faffing around with string->bytes wrappers. + bytes_io = BytesIO(json_bytes) + + producer = NoRangeStaticProducer(request, bytes_io) + producer.start() return NOT_DONE_YET diff --git a/synapse/http/servlet.py b/synapse/http/servlet.py index 69f7085291..a1e4b88e6d 100644 --- a/synapse/http/servlet.py +++ b/synapse/http/servlet.py @@ -29,7 +29,7 @@ def parse_integer(request, name, default=None, required=False): Args: request: the twisted HTTP request. - name (str): the name of the query parameter. + name (bytes/unicode): the name of the query parameter. default (int|None): value to use if the parameter is absent, defaults to None. required (bool): whether to raise a 400 SynapseError if the @@ -46,6 +46,10 @@ def parse_integer(request, name, default=None, required=False): def parse_integer_from_args(args, name, default=None, required=False): + + if not isinstance(name, bytes): + name = name.encode('ascii') + if name in args: try: return int(args[name][0]) @@ -65,7 +69,7 @@ def parse_boolean(request, name, default=None, required=False): Args: request: the twisted HTTP request. - name (str): the name of the query parameter. + name (bytes/unicode): the name of the query parameter. default (bool|None): value to use if the parameter is absent, defaults to None. required (bool): whether to raise a 400 SynapseError if the @@ -83,11 +87,15 @@ def parse_boolean(request, name, default=None, required=False): def parse_boolean_from_args(args, name, default=None, required=False): + + if not isinstance(name, bytes): + name = name.encode('ascii') + if name in args: try: return { - "true": True, - "false": False, + b"true": True, + b"false": False, }[args[name][0]] except Exception: message = ( @@ -104,21 +112,29 @@ def parse_boolean_from_args(args, name, default=None, required=False): def parse_string(request, name, default=None, required=False, - allowed_values=None, param_type="string"): - """Parse a string parameter from the request query string. + allowed_values=None, param_type="string", encoding='ascii'): + """ + Parse a string parameter from the request query string. + + If encoding is not None, the content of the query param will be + decoded to Unicode using the encoding, otherwise it will be encoded Args: request: the twisted HTTP request. - name (str): the name of the query parameter. - default (str|None): value to use if the parameter is absent, defaults - to None. + name (bytes/unicode): the name of the query parameter. + default (bytes/unicode|None): value to use if the parameter is absent, + defaults to None. Must be bytes if encoding is None. required (bool): whether to raise a 400 SynapseError if the parameter is absent, defaults to False. - allowed_values (list[str]): List of allowed values for the string, - or None if any value is allowed, defaults to None + allowed_values (list[bytes/unicode]): List of allowed values for the + string, or None if any value is allowed, defaults to None. Must be + the same type as name, if given. + encoding: The encoding to decode the name to, and decode the string + content with. Returns: - str|None: A string value or the default. + bytes/unicode|None: A string value or the default. Unicode if encoding + was given, bytes otherwise. Raises: SynapseError if the parameter is absent and required, or if the @@ -126,14 +142,22 @@ def parse_string(request, name, default=None, required=False, is not one of those allowed values. """ return parse_string_from_args( - request.args, name, default, required, allowed_values, param_type, + request.args, name, default, required, allowed_values, param_type, encoding ) def parse_string_from_args(args, name, default=None, required=False, - allowed_values=None, param_type="string"): + allowed_values=None, param_type="string", encoding='ascii'): + + if not isinstance(name, bytes): + name = name.encode('ascii') + if name in args: value = args[name][0] + + if encoding: + value = value.decode(encoding) + if allowed_values is not None and value not in allowed_values: message = "Query parameter %r must be one of [%s]" % ( name, ", ".join(repr(v) for v in allowed_values) @@ -146,6 +170,10 @@ def parse_string_from_args(args, name, default=None, required=False, message = "Missing %s query parameter %r" % (param_type, name) raise SynapseError(400, message, errcode=Codes.MISSING_PARAM) else: + + if encoding and isinstance(default, bytes): + return default.decode(encoding) + return default diff --git a/synapse/http/site.py b/synapse/http/site.py index 5fd30a4c2c..88ed3714f9 100644 --- a/synapse/http/site.py +++ b/synapse/http/site.py @@ -11,7 +11,6 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import contextlib import logging import time @@ -19,8 +18,8 @@ import time from twisted.web.server import Request, Site from synapse.http import redact_uri -from synapse.http.request_metrics import RequestMetrics -from synapse.util.logcontext import ContextResourceUsage, LoggingContext +from synapse.http.request_metrics import RequestMetrics, requests_counter +from synapse.util.logcontext import LoggingContext, PreserveLoggingContext logger = logging.getLogger(__name__) @@ -34,25 +33,43 @@ class SynapseRequest(Request): It extends twisted's twisted.web.server.Request, and adds: * Unique request ID + * A log context associated with the request * Redaction of access_token query-params in __repr__ * Logging at start and end * Metrics to record CPU, wallclock and DB time by endpoint. - It provides a method `processing` which should be called by the Resource - which is handling the request, and returns a context manager. + It also provides a method `processing`, which returns a context manager. If this + method is called, the request won't be logged until the context manager is closed; + this is useful for asynchronous request handlers which may go on processing the + request even after the client has disconnected. + Attributes: + logcontext(LoggingContext) : the log context for this request """ def __init__(self, site, channel, *args, **kw): Request.__init__(self, channel, *args, **kw) self.site = site - self._channel = channel + self._channel = channel # this is used by the tests self.authenticated_entity = None self.start_time = 0 + # we can't yet create the logcontext, as we don't know the method. + self.logcontext = None + global _next_request_seq self.request_seq = _next_request_seq _next_request_seq += 1 + # whether an asynchronous request handler has called processing() + self._is_processing = False + + # the time when the asynchronous request handler completed its processing + self._processing_finished_time = None + + # what time we finished sending the response to the client (or the connection + # dropped) + self.finish_time = None + def __repr__(self): # We overwrite this so that we don't log ``access_token`` return '<%s at 0x%x method=%r uri=%r clientproto=%r site=%r>' % ( @@ -74,11 +91,116 @@ class SynapseRequest(Request): return self.requestHeaders.getRawHeaders(b"User-Agent", [None])[-1] def render(self, resrc): + # this is called once a Resource has been found to serve the request; in our + # case the Resource in question will normally be a JsonResource. + + # create a LogContext for this request + request_id = self.get_request_id() + logcontext = self.logcontext = LoggingContext(request_id) + logcontext.request = request_id + # override the Server header which is set by twisted self.setHeader("Server", self.site.server_version_string) - return Request.render(self, resrc) + + with PreserveLoggingContext(self.logcontext): + # we start the request metrics timer here with an initial stab + # at the servlet name. For most requests that name will be + # JsonResource (or a subclass), and JsonResource._async_render + # will update it once it picks a servlet. + servlet_name = resrc.__class__.__name__ + self._started_processing(servlet_name) + + Request.render(self, resrc) + + # record the arrival of the request *after* + # dispatching to the handler, so that the handler + # can update the servlet name in the request + # metrics + requests_counter.labels(self.method, + self.request_metrics.name).inc() + + @contextlib.contextmanager + def processing(self): + """Record the fact that we are processing this request. + + Returns a context manager; the correct way to use this is: + + @defer.inlineCallbacks + def handle_request(request): + with request.processing("FooServlet"): + yield really_handle_the_request() + + Once the context manager is closed, the completion of the request will be logged, + and the various metrics will be updated. + """ + if self._is_processing: + raise RuntimeError("Request is already processing") + self._is_processing = True + + try: + yield + except Exception: + # this should already have been caught, and sent back to the client as a 500. + logger.exception("Asynchronous messge handler raised an uncaught exception") + finally: + # the request handler has finished its work and either sent the whole response + # back, or handed over responsibility to a Producer. + + self._processing_finished_time = time.time() + self._is_processing = False + + # if we've already sent the response, log it now; otherwise, we wait for the + # response to be sent. + if self.finish_time is not None: + self._finished_processing() + + def finish(self): + """Called when all response data has been written to this Request. + + Overrides twisted.web.server.Request.finish to record the finish time and do + logging. + """ + self.finish_time = time.time() + Request.finish(self) + if not self._is_processing: + with PreserveLoggingContext(self.logcontext): + self._finished_processing() + + def connectionLost(self, reason): + """Called when the client connection is closed before the response is written. + + Overrides twisted.web.server.Request.connectionLost to record the finish time and + do logging. + """ + self.finish_time = time.time() + Request.connectionLost(self, reason) + + # we only get here if the connection to the client drops before we send + # the response. + # + # It's useful to log it here so that we can get an idea of when + # the client disconnects. + with PreserveLoggingContext(self.logcontext): + logger.warn( + "Error processing request %r: %s %s", self, reason.type, reason.value, + ) + + if not self._is_processing: + self._finished_processing() def _started_processing(self, servlet_name): + """Record the fact that we are processing this request. + + This will log the request's arrival. Once the request completes, + be sure to call finished_processing. + + Args: + servlet_name (str): the name of the servlet which will be + processing this request. This is used in the metrics. + + It is possible to update this afterwards by updating + self.request_metrics.name. + """ self.start_time = time.time() self.request_metrics = RequestMetrics() self.request_metrics.start( @@ -94,18 +216,32 @@ class SynapseRequest(Request): ) def _finished_processing(self): - try: - context = LoggingContext.current_context() - usage = context.get_resource_usage() - except Exception: - usage = ContextResourceUsage() + """Log the completion of this request and update the metrics + """ + + if self.logcontext is None: + # this can happen if the connection closed before we read the + # headers (so render was never called). In that case we'll already + # have logged a warning, so just bail out. + return + + usage = self.logcontext.get_resource_usage() + + if self._processing_finished_time is None: + # we completed the request without anything calling processing() + self._processing_finished_time = time.time() - end_time = time.time() + # the time between receiving the request and the request handler finishing + processing_time = self._processing_finished_time - self.start_time + + # the time between the request handler finishing and the response being sent + # to the client (nb may be negative) + response_send_time = self.finish_time - self._processing_finished_time # need to decode as it could be raw utf-8 bytes # from a IDN servname in an auth header authenticated_entity = self.authenticated_entity - if authenticated_entity is not None: + if authenticated_entity is not None and isinstance(authenticated_entity, bytes): authenticated_entity = authenticated_entity.decode("utf-8", "replace") # ...or could be raw utf-8 bytes in the User-Agent header. @@ -116,22 +252,31 @@ class SynapseRequest(Request): user_agent = self.get_user_agent() if user_agent is not None: user_agent = user_agent.decode("utf-8", "replace") + else: + user_agent = "-" + + code = str(self.code) + if not self.finished: + # we didn't send the full response before we gave up (presumably because + # the connection dropped) + code += "!" self.site.access_logger.info( "%s - %s - {%s}" - " Processed request: %.3fsec (%.3fsec, %.3fsec) (%.3fsec/%.3fsec/%d)" + " Processed request: %.3fsec/%.3fsec (%.3fsec, %.3fsec) (%.3fsec/%.3fsec/%d)" " %sB %s \"%s %s %s\" \"%s\" [%d dbevts]", self.getClientIP(), self.site.site_tag, authenticated_entity, - end_time - self.start_time, + processing_time, + response_send_time, usage.ru_utime, usage.ru_stime, usage.db_sched_duration_sec, usage.db_txn_duration_sec, int(usage.db_txn_count), self.sentLength, - self.code, + code, self.method, self.get_redacted_uri(), self.clientproto, @@ -140,38 +285,10 @@ class SynapseRequest(Request): ) try: - self.request_metrics.stop(end_time, self) + self.request_metrics.stop(self.finish_time, self) except Exception as e: logger.warn("Failed to stop metrics: %r", e) - @contextlib.contextmanager - def processing(self, servlet_name): - """Record the fact that we are processing this request. - - Returns a context manager; the correct way to use this is: - - @defer.inlineCallbacks - def handle_request(request): - with request.processing("FooServlet"): - yield really_handle_the_request() - - This will log the request's arrival. Once the context manager is - closed, the completion of the request will be logged, and the various - metrics will be updated. - - Args: - servlet_name (str): the name of the servlet which will be - processing this request. This is used in the metrics. - - It is possible to update this afterwards by updating - self.request_metrics.servlet_name. - """ - # TODO: we should probably just move this into render() and finish(), - # to save having to call a separate method. - self._started_processing(servlet_name) - yield - self._finished_processing() - class XForwardedForRequest(SynapseRequest): def __init__(self, *args, **kw): @@ -217,7 +334,7 @@ class SynapseSite(Site): proxied = config.get("x_forwarded", False) self.requestFactory = SynapseRequestFactory(self, proxied) self.access_logger = logging.getLogger(logger_name) - self.server_version_string = server_version_string + self.server_version_string = server_version_string.encode('ascii') def log(self, request): pass |