Resource tracking for background processes

This introduces a mechanism for tracking resource usage by background
processes, along with an example of how it will be used.

This will help address #3518, but more importantly will give us better insights
into things which are happening but not being shown up by the request metrics.

We *could* do this with Measure blocks, but:
 - I think having them pulled out as a completely separate metric class will
   make it easier to distinguish top-level processes from those which are
   nested.

 - I want to be able to report on in-flight background processes, and I don't
   think we want to do this for *all* Measure blocks.

1 files changed, 7 insertions, 5 deletions

diff --git a/synapse/federation/transaction_queue.py b/synapse/federation/transaction_queue.py

index 5a956ecfb3..5c5a73b73c 100644
--- a/synapse/federation/transaction_queue.py
+++ b/synapse/federation/transaction_queue.py
@@ -30,7 +30,8 @@ from synapse.metrics import (
     sent_edus_counter,
     sent_transactions_counter,
 )
-from synapse.util import PreserveLoggingContext, logcontext
+from synapse.metrics.background_process_metrics import run_as_background_process
+from synapse.util import logcontext
 from synapse.util.metrics import measure_func
 from synapse.util.retryutils import NotRetryingDestination, get_retry_limiter
 
@@ -165,10 +166,11 @@ class TransactionQueue(object):
         if self._is_processing:
             return
 
-        # fire off a processing loop in the background. It's likely it will
-        # outlast the current request, so run it in the sentinel logcontext.
-        with PreserveLoggingContext():
-            self._process_event_queue_loop()
+        # fire off a processing loop in the background
+        run_as_background_process(
+            "process_transaction_queue",
+            self._process_event_queue_loop,
+        )
 
     @defer.inlineCallbacks
     def _process_event_queue_loop(self):