service-capacity-modeling 0.3.89__py3-none-any.whl → 0.3.90__py3-none-any.whl

service_capacity_modeling/models/org/netflix/cassandra.py

@@ -638,8 +638,14 @@ def _target_rf(desires: CapacityDesires, user_copies: Optional[int]) -> int:
 
 
 class NflxCassandraArguments(BaseModel):
-    copies_per_region: int = Field(
-        default=3,
+    """Configuration arguments for the Netflix Cassandra capacity model.
+
+    This model centralizes all tunable parameters with their defaults.
+    Use `from_extra_model_arguments()` to parse a dict into a validated instance.
+    """
+
+    copies_per_region: Optional[int] = Field(
+        default=None,
         description="How many copies of the data will exist e.g. RF=3. If unsupplied"
         " this will be deduced from durability and consistency desires",
     )
@@ -663,9 +669,13 @@ class NflxCassandraArguments(BaseModel):
         default=192,
         description="What is the maximum size of a cluster in this region",
     )
-    max_local_disk_gib: int = Field(
-        default=5120,
-        description="The maximum amount of data we store per machine",
+    max_local_data_per_node_gib: int = Field(
+        default=1280,
+        description="Maximum data per node for local disk instances (GiB)",
+    )
+    max_attached_data_per_node_gib: int = Field(
+        default=2048,
+        description="Maximum data per node for attached disk instances (GiB)",
     )
     max_write_buffer_percent: float = Field(
         default=0.25,
@@ -680,6 +690,26 @@ class NflxCassandraArguments(BaseModel):
         "automatically adjust to 0.2",
     )
 
+    @classmethod
+    def from_extra_model_arguments(
+        cls, extra_model_arguments: Dict[str, Any]
+    ) -> "NflxCassandraArguments":
+        """Parse extra_model_arguments dict into a validated NflxCassandraArguments.
+
+        This centralizes default values - any field not in extra_model_arguments
+        will use the default defined in this model.
+
+        Handles legacy field name mappings:
+        - max_local_disk_gib -> max_local_data_per_node_gib (if not explicitly set)
+        """
+        # Handle legacy field name: max_local_disk_gib -> max_local_data_per_node_gib
+        args = dict(extra_model_arguments)
+        if "max_local_data_per_node_gib" not in args and "max_local_disk_gib" in args:
+            args["max_local_data_per_node_gib"] = args["max_local_disk_gib"]
+
+        # Pydantic will use defaults for any missing fields
+        return cls.model_validate(args)
+
 
 class NflxCassandraCapacityModel(CapacityModel):
     def __init__(self) -> None:
@@ -722,40 +752,22 @@ class NflxCassandraCapacityModel(CapacityModel):
         desires: CapacityDesires,
         extra_model_arguments: Dict[str, Any],
     ) -> Optional[CapacityPlan]:
-        # TODO: Standardize these extra model argument defaults in a single
-        # place. Many of them are defined here and as default values in the
-        # downstream method but only these ones are used which is confusing for
-        # readability
-
-        # Use durabiliy and consistency to compute RF.
-        copies_per_region = _target_rf(
-            desires, extra_model_arguments.get("copies_per_region", None)
-        )
-        require_local_disks: bool = extra_model_arguments.get(
-            "require_local_disks", False
-        )
-        require_attached_disks: bool = extra_model_arguments.get(
-            "require_attached_disks", False
-        )
+        # Parse extra_model_arguments into a validated model with centralized defaults
+        args = NflxCassandraArguments.from_extra_model_arguments(extra_model_arguments)
+
+        # Use durability and consistency to compute RF if not explicitly set
+        copies_per_region = _target_rf(desires, args.copies_per_region)
+
+        # Validate required_cluster_size for critical tiers
         required_cluster_size: Optional[int] = (
             NflxCassandraCapacityModel.get_required_cluster_size(
                 desires.service_tier, extra_model_arguments
             )
         )
 
-        max_rps_to_disk: int = extra_model_arguments.get("max_rps_to_disk", 500)
-        max_regional_size: int = extra_model_arguments.get("max_regional_size", 192)
-        max_local_data_per_node_gib: int = extra_model_arguments.get(
-            "max_local_data_per_node_gib",
-            extra_model_arguments.get("max_local_disk_gib", 1280),
-        )
-
-        max_write_buffer_percent: float = min(
-            0.5, extra_model_arguments.get("max_write_buffer_percent", 0.25)
-        )
-        max_table_buffer_percent: float = min(
-            0.5, extra_model_arguments.get("max_table_buffer_percent", 0.11)
-        )
+        # Apply caps to buffer percentages
+        max_write_buffer_percent = min(0.5, args.max_write_buffer_percent)
+        max_table_buffer_percent = min(0.5, args.max_table_buffer_percent)
 
         # Adjust heap defaults for high write clusters
         if (
@@ -772,12 +784,12 @@ class NflxCassandraCapacityModel(CapacityModel):
             desires=desires,
             zones_per_region=context.zones_in_region,
             copies_per_region=copies_per_region,
-            require_local_disks=require_local_disks,
-            require_attached_disks=require_attached_disks,
+            require_local_disks=args.require_local_disks,
+            require_attached_disks=args.require_attached_disks,
             required_cluster_size=required_cluster_size,
-            max_rps_to_disk=max_rps_to_disk,
-            max_regional_size=max_regional_size,
-            max_local_data_per_node_gib=max_local_data_per_node_gib,
+            max_rps_to_disk=args.max_rps_to_disk,
+            max_regional_size=args.max_regional_size,
+            max_local_data_per_node_gib=args.max_local_data_per_node_gib,
             max_write_buffer_percent=max_write_buffer_percent,
             max_table_buffer_percent=max_table_buffer_percent,
         )

{service_capacity_modeling-0.3.89.dist-info → service_capacity_modeling-0.3.90.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: service-capacity-modeling
-Version: 0.3.89
+Version: 0.3.90
 Summary: Contains utilities for modeling capacity for pluggable workloads
 Author: Joseph Lynch
 Author-email: josephl@netflix.com

{service_capacity_modeling-0.3.89.dist-info → service_capacity_modeling-0.3.90.dist-info}/RECORD

@@ -53,7 +53,7 @@ service_capacity_modeling/models/utils.py,sha256=WosEEg4o1_WSbTb5mL-M1v8JuWJgvS2
 service_capacity_modeling/models/org/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 service_capacity_modeling/models/org/netflix/__init__.py,sha256=2Ld2NPxiO3vbYtOMqHtrV4f1nEH390Hoxqo3I5NbBDI,2553
 service_capacity_modeling/models/org/netflix/aurora.py,sha256=Js33ZjxCtt34HiDPsWRT9mjKCAsnnCo9du15QArVFMo,13073
-service_capacity_modeling/models/org/netflix/cassandra.py,sha256=Fp37bHtWRJctJYYJhW78YbLeFXoX26QxREP_BMEItNE,39003
+service_capacity_modeling/models/org/netflix/cassandra.py,sha256=MUILIhdxlqIxOwMbFJvy7luRYqBxraIkIVU_ZKnjxOo,39627
 service_capacity_modeling/models/org/netflix/control.py,sha256=wkJnqG7Nn7kQNcKEMzl9LFlA9xTAji2wYxL011VAKPI,5762
 service_capacity_modeling/models/org/netflix/counter.py,sha256=kTDL7dCnkn-XU27_Z1VBc4CCLCPoOqJZe9WgcENHHd4,10517
 service_capacity_modeling/models/org/netflix/crdb.py,sha256=iW7tyG8jpXhHIdXrw3DPYSHRAknPN42MlCRLJO4o9C8,20826
@@ -77,9 +77,9 @@ service_capacity_modeling/tools/auto_shape.py,sha256=Jx9H2ay9-H_kUDjtB141owQNxGF
 service_capacity_modeling/tools/fetch_pricing.py,sha256=Qp-XMymkY1dvtyS51RufmEpfgOHv-IQ-XyzS8wp2-qM,4021
 service_capacity_modeling/tools/generate_missing.py,sha256=F7YqvMJAV4nZc20GNrlIsnQSF8_77sLgwYZqc5k4LDg,3099
 service_capacity_modeling/tools/instance_families.py,sha256=e5RuYkCLUITvsAazDH12B6KjX_PaBsv6Ne3mj0HK_sQ,9223
-service_capacity_modeling-0.3.89.dist-info/licenses/LICENSE,sha256=nl_Lt5v9VvJ-5lWJDT4ddKAG-VZ-2IaLmbzpgYDz2hU,11343
-service_capacity_modeling-0.3.89.dist-info/METADATA,sha256=ATxeW6sDZDkk4HQpmYGdmIlAXVXYj7zQaK5QJwkml8o,10366
-service_capacity_modeling-0.3.89.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-service_capacity_modeling-0.3.89.dist-info/entry_points.txt,sha256=ZsjzpG5SomWpT1zCE19n1uSXKH2gTI_yc33sdl0vmJg,146
-service_capacity_modeling-0.3.89.dist-info/top_level.txt,sha256=H8XjTCLgR3enHq5t3bIbxt9SeUkUT8HT_SDv2dgIT_A,26
-service_capacity_modeling-0.3.89.dist-info/RECORD,,
+service_capacity_modeling-0.3.90.dist-info/licenses/LICENSE,sha256=nl_Lt5v9VvJ-5lWJDT4ddKAG-VZ-2IaLmbzpgYDz2hU,11343
+service_capacity_modeling-0.3.90.dist-info/METADATA,sha256=QAmQcpV4CC-6rKvYKnmzrb4OjpW7sCQdaFAdB5u9TVo,10366
+service_capacity_modeling-0.3.90.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+service_capacity_modeling-0.3.90.dist-info/entry_points.txt,sha256=ZsjzpG5SomWpT1zCE19n1uSXKH2gTI_yc33sdl0vmJg,146
+service_capacity_modeling-0.3.90.dist-info/top_level.txt,sha256=H8XjTCLgR3enHq5t3bIbxt9SeUkUT8HT_SDv2dgIT_A,26
+service_capacity_modeling-0.3.90.dist-info/RECORD,,