diff --git a/lib/carbon/cache.py b/lib/carbon/cache.py
index cf8b7d1e..72ccf6cf 100644
--- a/lib/carbon/cache.py
+++ b/lib/carbon/cache.py
@@ -16,7 +16,7 @@
 import threading
 from operator import itemgetter
 from random import choice
-from collections import defaultdict
+from collections import defaultdict, deque
 
 from carbon.conf import settings
 from carbon import events, log
@@ -189,6 +189,7 @@ class _MetricCache(defaultdict):
   def __init__(self, strategy=None):
     self.lock = threading.Lock()
     self.size = 0
+    self.new_metrics = deque()
     self.strategy = None
     if strategy:
       self.strategy = strategy(self)
@@ -253,6 +254,8 @@ def store(self, metric, datapoint):
           log.msg("MetricCache is full: self.size=%d" % self.size)
           events.cacheFull()
         else:
+          if not self[metric]:
+            self.new_metrics.append(metric)
           self.size += 1
           self[metric][timestamp] = value
           if self.strategy:
diff --git a/lib/carbon/writer.py b/lib/carbon/writer.py
index 9161d95a..adcc0a01 100644
--- a/lib/carbon/writer.py
+++ b/lib/carbon/writer.py
@@ -95,24 +95,25 @@ def writeCachedDataPoints():
 
   cache = MetricCache()
   while cache:
-    (metric, datapoints) = cache.drain_metric()
-    if metric is None:
-      # end the loop
-      break
+    # First, create new metrics files, which is helpful for graphite-web
+    while cache.new_metrics and (not CREATE_BUCKET or CREATE_BUCKET.peek(1)):
+      metric = cache.new_metrics.popleft()
 
-    dbFileExists = state.database.exists(metric)
+      if metric not in cache:
+        # This metric has already been drained. There's no sense in creating it.
+        continue
 
-    if not dbFileExists:
-      if CREATE_BUCKET and not CREATE_BUCKET.drain(1):
-        # If our tokenbucket doesn't have enough tokens available to create a new metric
-        # file then we'll just drop the metric on the ground and move on to the next
-        # metric.
-        # XXX This behavior should probably be configurable to no tdrop metrics
-        # when rate limiting unless our cache is too big or some other legit
-        # reason.
-        instrumentation.increment('droppedCreates')
+      if state.database.exists(metric):
         continue
 
+      if CREATE_BUCKET and not CREATE_BUCKET.drain(1):
+        # This should never actually happen as no other thread should be
+        # draining our tokens, and we just checked for a token.
+        # Just put the new metric back in the create list and we'll try again
+        # after writing an update.
+        cache.new_metrics.appendleft(metric)
+        break
+
       archiveConfig = None
       xFilesFactor, aggregationMethod = None, None
 
@@ -150,6 +151,18 @@ def writeCachedDataPoints():
         instrumentation.increment('errors')
         continue
 
+    # now drain and persist some data
+    (metric, datapoints) = cache.drain_metric()
+    if metric is None:
+      # end the loop
+      break
+
+    if not state.database.exists(metric):
+      # If we get here, the metric must still be in new_metrics. We're
+      # creating too fast, and we'll drop this data.
+      instrumentation.increment('droppedCreates')
+      continue
+
     # If we've got a rate limit configured lets makes sure we enforce it
     waitTime = 0
     if UPDATE_BUCKET: