Instrument `FederationStateIdsServlet` - `/state_ids` (#13499)

Instrument FederationStateIdsServlet - `/state_ids` so it's easier to follow what's going on in Jaeger when viewing a trace.

5 files changed, 21 insertions, 2 deletions

diff --git a/changelog.d/13499.misc b/changelog.d/13499.misc
new file mode 100644
index 0000000000..99dbcebec8
--- /dev/null
+++ b/changelog.d/13499.misc
@@ -0,0 +1 @@
+Instrument `FederationStateIdsServlet` (`/state_ids`) for understandable traces in Jaeger.
diff --git a/synapse/federation/federation_server.py b/synapse/federation/federation_server.py
index db4b83a505..75fbc6073d 100644
--- a/synapse/federation/federation_server.py
+++ b/synapse/federation/federation_server.py
@@ -61,7 +61,12 @@ from synapse.logging.context import (
     nested_logging_context,
     run_in_background,
 )
-from synapse.logging.opentracing import log_kv, start_active_span_from_edu, trace
+from synapse.logging.opentracing import (
+    log_kv,
+    start_active_span_from_edu,
+    tag_args,
+    trace,
+)
 from synapse.metrics.background_process_metrics import wrap_as_background_process
 from synapse.replication.http.federation import (
     ReplicationFederationSendEduRestServlet,
@@ -547,6 +552,8 @@ class FederationServer(FederationBase):
 
         return 200, resp
 
+    @trace
+    @tag_args
     async def on_state_ids_request(
         self, origin: str, room_id: str, event_id: str
     ) -> Tuple[int, JsonDict]:
@@ -569,6 +576,8 @@ class FederationServer(FederationBase):
 
         return 200, resp
 
+    @trace
+    @tag_args
     async def _on_state_ids_request_compute(
         self, room_id: str, event_id: str
     ) -> JsonDict:
diff --git a/synapse/handlers/federation.py b/synapse/handlers/federation.py
index 5042236742..6f5ab86ac4 100644
--- a/synapse/handlers/federation.py
+++ b/synapse/handlers/federation.py
@@ -59,7 +59,7 @@ from synapse.events.validator import EventValidator
 from synapse.federation.federation_client import InvalidResponseError
 from synapse.http.servlet import assert_params_in_dict
 from synapse.logging.context import nested_logging_context
-from synapse.logging.opentracing import trace
+from synapse.logging.opentracing import tag_args, trace
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.module_api import NOT_SPAM
 from synapse.replication.http.federation import (
@@ -1081,6 +1081,8 @@ class FederationHandler:
 
         return event
 
+    @trace
+    @tag_args
     async def get_state_ids_for_pdu(self, room_id: str, event_id: str) -> List[str]:
         """Returns the state at the event. i.e. not including said event."""
         event = await self.store.get_event(event_id, check_room_id=room_id)
diff --git a/synapse/storage/databases/main/event_federation.py b/synapse/storage/databases/main/event_federation.py
index eec55b6478..0bc8401f2b 100644
--- a/synapse/storage/databases/main/event_federation.py
+++ b/synapse/storage/databases/main/event_federation.py
@@ -33,6 +33,7 @@ from synapse.api.constants import MAX_DEPTH, EventTypes
 from synapse.api.errors import StoreError
 from synapse.api.room_versions import EventFormatVersions, RoomVersion
 from synapse.events import EventBase, make_event_from_dict
+from synapse.logging.opentracing import tag_args, trace
 from synapse.metrics.background_process_metrics import wrap_as_background_process
 from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause
 from synapse.storage.database import (
@@ -126,6 +127,8 @@ class EventFederationWorkerStore(SignatureWorkerStore, EventsWorkerStore, SQLBas
         )
         return await self.get_events_as_list(event_ids)
 
+    @trace
+    @tag_args
     async def get_auth_chain_ids(
         self,
         room_id: str,
diff --git a/synapse/util/ratelimitutils.py b/synapse/util/ratelimitutils.py
index 6394cc39ac..e1beaec5a3 100644
--- a/synapse/util/ratelimitutils.py
+++ b/synapse/util/ratelimitutils.py
@@ -27,6 +27,7 @@ from synapse.logging.context import (
     make_deferred_yieldable,
     run_in_background,
 )
+from synapse.logging.opentracing import start_active_span
 from synapse.util import Clock
 
 if typing.TYPE_CHECKING:
@@ -176,8 +177,11 @@ class _PerHostRatelimiter:
             # Ensure that we've properly cleaned up.
             self.sleeping_requests.discard(request_id)
             self.ready_request_queue.pop(request_id, None)
+            wait_span_scope.__exit__(None, None, None)
             return r
 
+        wait_span_scope = start_active_span("ratelimit wait")
+        wait_span_scope.__enter__()
         ret_defer.addCallbacks(on_start, on_err)
         ret_defer.addBoth(on_both)
         return make_deferred_yieldable(ret_defer)