Merge pull request #121 from Netflix-Skunkworks/feature/use-buffers

support derived buffer passed in the input desires

service_capacity_modeling/interface.py

@@ -897,9 +897,10 @@ def merge_with(self, defaults: "CapacityDesires") -> "CapacityDesires":
             default_buffers["default"] = desired_buffers["default"]
         for k, v in desired_buffers.get("desired", {}).items():
             default_buffers["desired"][k] = v
-        for i in desired_buffers.get("derived", []):
-            if i not in default_buffers["derived"]:
-                default_buffers["derived"].append(i)
+
+        default_buffers.setdefault("derived", {})
+        for k, v in desired_buffers.get("derived", {}).items():
+            default_buffers["derived"][k] = v
 
         default_dict.update(desires_dict)
 

tests/test_desire_merge.py

@@ -1,6 +1,7 @@
 from service_capacity_modeling.capacity_planner import planner
 from service_capacity_modeling.interface import Buffer
 from service_capacity_modeling.interface import BufferComponent
+from service_capacity_modeling.interface import BufferIntent
 from service_capacity_modeling.interface import Buffers
 from service_capacity_modeling.interface import CapacityDesires
 from service_capacity_modeling.interface import certain_int
@@ -25,7 +26,12 @@
         desired={
             "custom": Buffer(ratio=3.8, components=["custom"]),
             "custom-cpu": Buffer(ratio=3.0, components=[BufferComponent.cpu]),
-        }
+        },
+        derived={
+            "compute": Buffer(
+                intent=BufferIntent.scale, ratio=2, components=["compute"]
+            )
+        },
     ),
 )
 
@@ -41,6 +47,9 @@ def test_cassandra_merge():
     assert merged.query_pattern.estimated_read_per_second.mid == 100000
     assert merged.query_pattern.estimated_mean_read_size_bytes.low == 10
     assert merged.data_shape.estimated_state_size_gib.mid == 10
+    assert merged.buffers.derived.get("compute") is not None
+    assert merged.buffers.derived["compute"].ratio == 2.0
+    assert merged.buffers.derived["compute"].intent == BufferIntent.scale
 
     # Should come from cassandra model
     assert merged.query_pattern.estimated_mean_read_latency_ms.mid == 2.0