Merge pull request #3933 from matrix-org/erikj/destination_retry_cache

 Add a five minute cache to get_destination_retry_timings

1 files changed, 22 insertions, 1 deletions

diff --git a/synapse/storage/transactions.py b/synapse/storage/transactions.py
index baf0379a68..ab54977a75 100644
--- a/synapse/storage/transactions.py
+++ b/synapse/storage/transactions.py
@@ -23,6 +23,7 @@ from canonicaljson import encode_canonical_json
 from twisted.internet import defer
 
 from synapse.metrics.background_process_metrics import run_as_background_process
+from synapse.util.caches.expiringcache import ExpiringCache
 
 from ._base import SQLBaseStore, db_to_json
 
@@ -49,6 +50,8 @@ _UpdateTransactionRow = namedtuple(
     )
 )
 
+SENTINEL = object()
+
 
 class TransactionStore(SQLBaseStore):
     """A collection of queries for handling PDUs.
@@ -59,6 +62,12 @@ class TransactionStore(SQLBaseStore):
 
         self._clock.looping_call(self._start_cleanup_transactions, 30 * 60 * 1000)
 
+        self._destination_retry_cache = ExpiringCache(
+            cache_name="get_destination_retry_timings",
+            clock=self._clock,
+            expiry_ms=5 * 60 * 1000,
+        )
+
     def get_received_txn_response(self, transaction_id, origin):
         """For an incoming transaction from a given origin, check if we have
         already responded to it. If so, return the response code and response
@@ -155,6 +164,7 @@ class TransactionStore(SQLBaseStore):
         """
         pass
 
+    @defer.inlineCallbacks
     def get_destination_retry_timings(self, destination):
         """Gets the current retry timings (if any) for a given destination.
 
@@ -165,10 +175,20 @@ class TransactionStore(SQLBaseStore):
             None if not retrying
             Otherwise a dict for the retry scheme
         """
-        return self.runInteraction(
+
+        result = self._destination_retry_cache.get(destination, SENTINEL)
+        if result is not SENTINEL:
+            defer.returnValue(result)
+
+        result = yield self.runInteraction(
             "get_destination_retry_timings",
             self._get_destination_retry_timings, destination)
 
+        # We don't hugely care about race conditions between getting and
+        # invalidating the cache, since we time out fairly quickly anyway.
+        self._destination_retry_cache[destination] = result
+        defer.returnValue(result)
+
     def _get_destination_retry_timings(self, txn, destination):
         result = self._simple_select_one_txn(
             txn,
@@ -196,6 +216,7 @@ class TransactionStore(SQLBaseStore):
             retry_interval (int) - how long until next retry in ms
         """
 
+        self._destination_retry_cache.pop(destination)
         return self.runInteraction(
             "set_destination_retry_timings",
             self._set_destination_retry_timings,