Avoid rapidly backing-off a server if we ignore the retry interval

changelog.d/5335.bugfix

@@ -0,0 +1 @@
+Fix a bug where we could rapidly mark a server as unreachable even though it was only down for a few minutes.

synapse/util/retryutils.py

@@ -46,8 +46,7 @@ class NotRetryingDestination(Exception):
 
 
 @defer.inlineCallbacks
-def get_retry_limiter(destination, clock, store, ignore_backoff=False,
-                      **kwargs):
+def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs):
     """For a given destination check if we have previously failed to
     send a request there and are waiting before retrying the destination.
     If we are not ready to retry the destination, this will raise a
@@ -60,8 +59,7 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False,
         clock (synapse.util.clock): timing source
         store (synapse.storage.transactions.TransactionStore): datastore
         ignore_backoff (bool): true to ignore the historical backoff data and
-            try the request anyway. We will still update the next
-            retry_interval on success/failure.
+            try the request anyway. We will still reset the retry_interval on success.
 
     Example usage:
 
@@ -75,13 +73,12 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False,
     """
     retry_last_ts, retry_interval = (0, 0)
 
-    retry_timings = yield store.get_destination_retry_timings(
-        destination
-    )
+    retry_timings = yield store.get_destination_retry_timings(destination)
 
     if retry_timings:
         retry_last_ts, retry_interval = (
-            retry_timings["retry_last_ts"], retry_timings["retry_interval"]
+            retry_timings["retry_last_ts"],
+            retry_timings["retry_interval"],
         )
 
         now = int(clock.time_msec())
@@ -93,22 +90,31 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False,
                 destination=destination,
             )
 
+    # if we are ignoring the backoff data, we should also not increment the backoff
+    # when we get another failure - otherwise a server can very quickly reach the
+    # maximum backoff even though it might only have been down briefly
+    backoff_on_failure = not ignore_backoff
+
     defer.returnValue(
         RetryDestinationLimiter(
-            destination,
-            clock,
-            store,
-            retry_interval,
-            **kwargs
+            destination, clock, store, retry_interval, backoff_on_failure, **kwargs
         )
     )
 
 
 class RetryDestinationLimiter(object):
-    def __init__(self, destination, clock, store, retry_interval,
+    def __init__(
+        self,
+        destination,
+        clock,
+        store,
+        retry_interval,
         min_retry_interval=10 * 60 * 1000,
         max_retry_interval=24 * 60 * 60 * 1000,
-                 multiplier_retry_interval=5, backoff_on_404=False):
+        multiplier_retry_interval=5,
+        backoff_on_404=False,
+        backoff_on_failure=True,
+    ):
         """Marks the destination as "down" if an exception is thrown in the
         context, except for CodeMessageException with code < 500.
 
@@ -128,6 +134,9 @@ class RetryDestinationLimiter(object):
             multiplier_retry_interval (int): The multiplier to use to increase
                 the retry interval after a failed request.
             backoff_on_404 (bool): Back off if we get a 404
+
+            backoff_on_failure (bool): set to False if we should not increase the
+                retry interval on a failure.
         """
         self.clock = clock
         self.store = store
@@ -138,6 +147,7 @@ class RetryDestinationLimiter(object):
         self.max_retry_interval = max_retry_interval
         self.multiplier_retry_interval = multiplier_retry_interval
         self.backoff_on_404 = backoff_on_404
+        self.backoff_on_failure = backoff_on_failure
 
     def __enter__(self):
         pass
@@ -173,10 +183,13 @@ class RetryDestinationLimiter(object):
             if not self.retry_interval:
                 return
 
-            logger.debug("Connection to %s was successful; clearing backoff",
-                         self.destination)
+            logger.debug(
+                "Connection to %s was successful; clearing backoff", self.destination
+            )
             retry_last_ts = 0
             self.retry_interval = 0
+        elif not self.backoff_on_failure:
+            return
         else:
             # We couldn't connect.
             if self.retry_interval:
@@ -190,7 +203,10 @@ class RetryDestinationLimiter(object):
 
             logger.info(
                 "Connection to %s was unsuccessful (%s(%s)); backoff now %i",
-                self.destination, exc_type, exc_val, self.retry_interval
+                self.destination,
+                exc_type,
+                exc_val,
+                self.retry_interval,
             )
             retry_last_ts = int(self.clock.time_msec())
 
@@ -201,9 +217,7 @@ class RetryDestinationLimiter(object):
                     self.destination, retry_last_ts, self.retry_interval
                 )
             except Exception:
-                logger.exception(
-                    "Failed to store destination_retry_timings",
-                )
+                logger.exception("Failed to store destination_retry_timings")
 
         # we deliberately do this in the background.
         synapse.util.logcontext.run_in_background(store_retry_timings)