indexify 0.3.22__py3-none-any.whl → 0.3.24__py3-none-any.whl

indexify/cli/cli.py

@@ -7,7 +7,6 @@ from tensorlake.utils.logging import (
 configure_logging_early()
 
 import os
-import re
 import shutil
 import sys
 from importlib.metadata import version
@@ -33,6 +32,8 @@ from indexify.executor.executor_flavor import ExecutorFlavor
 from indexify.executor.function_executor.server.subprocess_function_executor_server_factory import (
     SubprocessFunctionExecutorServerFactory,
 )
+from indexify.executor.host_resources.host_resources import HostResourcesProvider
+from indexify.executor.host_resources.nvidia_gpu_allocator import NvidiaGPUAllocator
 from indexify.executor.monitoring.health_checker.generic_health_checker import (
     GenericHealthChecker,
 )
@@ -225,6 +226,7 @@ def executor(
         monitoring_server_port=monitoring_server_port,
         enable_grpc_state_reconciler=enable_grpc_state_reconciler,
         blob_store=blob_store,
+        host_resources_provider=HostResourcesProvider(NvidiaGPUAllocator(logger)),
     ).run()
 
 

indexify/executor/api_objects.py

@@ -10,6 +10,25 @@ class DataPayload(BaseModel):
     content_type: Optional[str] = None
 
 
+class NodeGPU(BaseModel):
+    count: int
+    model: str
+
+
+class TaskResources(BaseModel):
+    cpus: float
+    memory_mb: int
+    ephemeral_disk_mb: int
+    gpu: Optional[NodeGPU] = None
+
+
+class TaskRetryPolicy(BaseModel):
+    max_retries: int
+    initial_delay_sec: float
+    max_delay_sec: float
+    delay_multiplier: float
+
+
 class Task(BaseModel):
     id: str
     namespace: str
@@ -27,6 +46,9 @@ class Task(BaseModel):
     input_payload: Optional[DataPayload] = None
     reducer_input_payload: Optional[DataPayload] = None
     output_payload_uri_prefix: Optional[str] = None
+    timeout: Optional[int] = None  # in seconds
+    resources: Optional[TaskResources] = None
+    retry_policy: Optional[TaskRetryPolicy] = None
 
 
 class FunctionURI(BaseModel):

indexify/executor/executor.py

@@ -24,6 +24,7 @@ from .function_executor.server.function_executor_server_factory import (
 from .grpc.channel_manager import ChannelManager
 from .grpc.state_reconciler import ExecutorStateReconciler
 from .grpc.state_reporter import ExecutorStateReporter
+from .host_resources.host_resources import HostResourcesProvider
 from .metrics.executor import (
     METRIC_TASKS_COMPLETED_OUTCOME_ALL,
     METRIC_TASKS_COMPLETED_OUTCOME_ERROR_CUSTOMER_CODE,
@@ -71,6 +72,7 @@ class Executor:
         monitoring_server_port: int,
         enable_grpc_state_reconciler: bool,
         blob_store: BLOBStore,
+        host_resources_provider: HostResourcesProvider,
     ):
         self._logger = structlog.get_logger(module=__name__)
         self._is_shutdown: bool = False
@@ -118,6 +120,7 @@ class Executor:
             function_allowlist=self._function_allowlist,
             function_executor_states=self._function_executor_states,
             channel_manager=self._channel_manager,
+            host_resources_provider=host_resources_provider,
             logger=self._logger,
         )
         self._state_reporter.update_executor_status(
@@ -171,6 +174,7 @@ class Executor:
         executor_info: Dict[str, str] = {
             "id": id,
             "dev_mode": str(development_mode),
+            "flavor": flavor.name,
             "version": version,
             "code_path": str(code_path),
             "server_addr": server_addr,
@@ -179,6 +183,8 @@ class Executor:
             "enable_grpc_state_reconciler": str(enable_grpc_state_reconciler),
             "hostname": gethostname(),
         }
+        for key, value in labels.items():
+            executor_info["label_" + key] = value
         executor_info.update(function_allowlist_to_info_dict(function_allowlist))
         metric_executor_info.info(executor_info)
 

indexify/executor/function_executor/server/function_executor_server_factory.py

@@ -24,6 +24,10 @@ class FunctionExecutorServerConfiguration:
     graph_version: str
     image_uri: Optional[str]
     secret_names: List[str]
+    cpu_ms_per_sec: Optional[int]
+    memory_bytes: Optional[int]
+    disk_bytes: Optional[int]
+    gpu_count: int
 
 
 class FunctionExecutorServerFactory:

indexify/executor/function_executor/single_task_runner.py

@@ -1,4 +1,5 @@
 from collections.abc import Awaitable, Callable
+from math import ceil
 from typing import Any, Optional
 
 import grpc
@@ -131,16 +132,33 @@ class SingleTaskRunner:
         self._function_executor_state.function_executor = FunctionExecutor(
             server_factory=self._function_executor_server_factory, logger=self._logger
         )
+        task: Task = self._task_input.task
         config: FunctionExecutorServerConfiguration = (
             FunctionExecutorServerConfiguration(
                 executor_id=self._executor_id,
                 function_executor_id=self._function_executor_state.id,
-                namespace=self._task_input.task.namespace,
-                image_uri=self._task_input.task.image_uri,
-                secret_names=self._task_input.task.secret_names or [],
-                graph_name=self._task_input.task.compute_graph,
-                graph_version=self._task_input.task.graph_version,
-                function_name=self._task_input.task.compute_fn,
+                namespace=task.namespace,
+                graph_name=task.compute_graph,
+                graph_version=task.graph_version,
+                function_name=task.compute_fn,
+                image_uri=task.image_uri,
+                secret_names=task.secret_names or [],
+                cpu_ms_per_sec=(
+                    None
+                    if task.resources.cpus is None
+                    else ceil(task.resources.cpus * 1000)
+                ),
+                memory_bytes=(
+                    None
+                    if task.resources.memory_mb is None
+                    else task.resources.memory_mb * 1024 * 1024
+                ),
+                disk_bytes=(
+                    None
+                    if task.resources.ephemeral_disk_mb is None
+                    else task.resources.ephemeral_disk_mb * 1024 * 1024
+                ),
+                gpu_count=0 if task.resources.gpu is None else task.resources.gpu.count,
             )
         )
         initialize_request: InitializeRequest = InitializeRequest(

indexify/executor/grpc/function_executor_controller.py

@@ -9,6 +9,7 @@ from tensorlake.function_executor.proto.message_validator import MessageValidato
 
 from indexify.proto.executor_api_pb2 import (
     FunctionExecutorDescription,
+    FunctionExecutorResources,
 )
 from indexify.proto.executor_api_pb2 import (
     FunctionExecutorStatus as FunctionExecutorStatusProto,
@@ -42,6 +43,13 @@ def validate_function_executor_description(
     # image_uri is optional.
     # secret_names can be empty.
     # resource_limits is optional.
+    # TODO: Make resources required after we migrate Server to them.
+    # validator.required_field("resources")
+    # validator = MessageValidator(function_executor_description.resources)
+    # validator.required_field("cpu_ms_per_sec")
+    # validator.required_field("memory_bytes")
+    # validator.required_field("disk_bytes")
+    # validator.required_field("gpu_count")
 
 
 def function_executor_logger(
@@ -333,14 +341,24 @@ async def _create_function_executor(
         executor_id=executor_id,
         function_executor_id=function_executor_description.id,
         namespace=function_executor_description.namespace,
-        image_uri=None,
-        secret_names=list(function_executor_description.secret_names),
         graph_name=function_executor_description.graph_name,
         graph_version=function_executor_description.graph_version,
         function_name=function_executor_description.function_name,
+        image_uri=None,
+        secret_names=list(function_executor_description.secret_names),
+        cpu_ms_per_sec=None,
+        memory_bytes=None,
+        disk_bytes=None,
+        gpu_count=0,
     )
     if function_executor_description.HasField("image_uri"):
         config.image_uri = function_executor_description.image_uri
+    if function_executor_description.HasField("resources"):
+        resources: FunctionExecutorResources = function_executor_description.resources
+        config.cpu_ms_per_sec = resources.cpu_ms_per_sec
+        config.memory_bytes = resources.memory_bytes
+        config.disk_bytes = resources.disk_bytes
+        config.gpu_count = resources.gpu_count
 
     initialize_request: InitializeRequest = InitializeRequest(
         namespace=function_executor_description.namespace,

indexify/executor/grpc/state_reporter.py

@@ -18,10 +18,12 @@ from indexify.proto.executor_api_pb2 import (
 from indexify.proto.executor_api_pb2 import (
     FunctionExecutorStatus as FunctionExecutorStatusProto,
 )
+from indexify.proto.executor_api_pb2 import GPUModel as GPUModelProto
 from indexify.proto.executor_api_pb2 import (
-    GPUModel,
     GPUResources,
-    HostResources,
+)
+from indexify.proto.executor_api_pb2 import HostResources as HostResourcesProto
+from indexify.proto.executor_api_pb2 import (
     ReportExecutorStateRequest,
 )
 from indexify.proto.executor_api_pb2_grpc import ExecutorAPIStub
@@ -33,6 +35,7 @@ from ..function_executor.function_executor_states_container import (
     FunctionExecutorStatesContainer,
 )
 from ..function_executor.function_executor_status import FunctionExecutorStatus
+from ..host_resources.host_resources import HostResources, HostResourcesProvider
 from ..runtime_probes import RuntimeProbes
 from .channel_manager import ChannelManager
 from .metrics.state_reporter import (
@@ -57,6 +60,7 @@ class ExecutorStateReporter:
         function_allowlist: Optional[List[FunctionURI]],
         function_executor_states: FunctionExecutorStatesContainer,
         channel_manager: ChannelManager,
+        host_resources_provider: HostResourcesProvider,
         logger: Any,
         reporting_interval_sec: int = _REPORTING_INTERVAL_SEC,
     ):
@@ -72,6 +76,9 @@ class ExecutorStateReporter:
         self._channel_manager = channel_manager
         self._logger: Any = logger.bind(module=__name__)
         self._reporting_interval_sec: int = reporting_interval_sec
+        self._total_host_resources: HostResourcesProto = _host_resources_to_proto(
+            host_resources_provider.total_resources(logger)
+        )
 
         self._is_shutdown: bool = False
         self._executor_status: ExecutorStatus = ExecutorStatus.EXECUTOR_STATUS_UNKNOWN
@@ -130,7 +137,9 @@ class ExecutorStateReporter:
                 flavor=_to_grpc_executor_flavor(self._flavor, self._logger),
                 version=self._version,
                 status=self._executor_status,
-                free_resources=await self._fetch_free_host_resources(),
+                # Server requires free_resources to be set but ignores its value for now.
+                free_resources=self._total_host_resources,
+                total_resources=self._total_host_resources,
                 allowed_functions=self._allowed_functions,
                 function_executor_states=await self._fetch_function_executor_states(),
                 labels=self._labels,
@@ -151,18 +160,6 @@ class ExecutorStateReporter:
         """
         self._is_shutdown = True
 
-    async def _fetch_free_host_resources(self) -> HostResources:
-        # TODO: Implement host resource metrics reporting.
-        return HostResources(
-            cpu_count=0,
-            memory_bytes=0,
-            disk_bytes=0,
-            gpu=GPUResources(
-                count=0,
-                model=GPUModel.GPU_MODEL_UNKNOWN,
-            ),
-        )
-
     async def _fetch_function_executor_states(self) -> List[FunctionExecutorStateProto]:
         states = []
 
@@ -264,3 +261,20 @@ def _state_hash(state: ExecutorState) -> str:
     hasher = hashlib.sha256(usedforsecurity=False)
     hasher.update(serialized_state)
     return hasher.hexdigest()
+
+
+def _host_resources_to_proto(host_resources: HostResources) -> HostResourcesProto:
+    proto = HostResourcesProto(
+        cpu_count=host_resources.cpu_count,
+        memory_bytes=host_resources.memory_mb * 1024 * 1024,
+        disk_bytes=host_resources.disk_mb * 1024 * 1024,
+    )
+    if len(host_resources.gpus) > 0:
+        proto.gpu = GPUResources(
+            count=len(host_resources.gpus),
+            deprecated_model=GPUModelProto.GPU_MODEL_UNKNOWN,  # TODO: Remove this field
+            model=host_resources.gpus[
+                0
+            ].model.value,  # All GPUs should have the same model
+        )
+    return proto

indexify/executor/host_resources/host_resources.py

@@ -0,0 +1,50 @@
+from typing import List
+
+from pydantic import BaseModel
+
+from .nvidia_gpu import NvidiaGPUInfo
+from .nvidia_gpu_allocator import NvidiaGPUAllocator
+
+
+class HostResources(BaseModel):
+    cpu_count: int
+    memory_mb: int
+    disk_mb: int
+    gpus: List[NvidiaGPUInfo]
+
+
+class HostResourcesProvider:
+    """
+    HostResourcesProvider is a class that provides information about the host resources.
+    """
+
+    def __init__(self, gpu_allocator: NvidiaGPUAllocator):
+        self._gpu_allocator: NvidiaGPUAllocator = gpu_allocator
+
+    def total_resources(self, logger) -> HostResources:
+        """Returns all hardware resources that exist at the host.
+
+        Raises Exception on error.
+        """
+        logger = logger.bind(module=__name__)
+
+        return HostResources(
+            cpu_count=0,  # TODO: Implement for Linux and MacOS hosts
+            memory_mb=0,  # TODO: Implement for Linux and MacOS hosts
+            disk_mb=0,  # TODO: Implement for Linux and MacOS hosts
+            gpus=self._gpu_allocator.list_all(),
+        )
+
+    def free_resources(self, logger) -> HostResources:
+        """Returns all hardware resources that are free at the host.
+
+        Raises Exception on error.
+        """
+        logger = logger.bind(module=__name__)
+
+        return HostResources(
+            cpu_count=0,  # TODO: Implement for Linux and MacOS hosts
+            memory_mb=0,  # TODO: Implement for Linux and MacOS hosts
+            disk_mb=0,  # TODO: Implement for Linux and MacOS hosts
+            gpus=self._gpu_allocator.list_free(),
+        )

indexify/executor/host_resources/nvidia_gpu.py

@@ -0,0 +1,77 @@
+import subprocess
+from enum import Enum
+from typing import Any, List
+
+from pydantic import BaseModel
+from tensorlake.functions_sdk.resources import GPU_MODEL
+
+
+# Only NVIDIA GPUs currently supported in Tensorlake SDK are listed here.
+class NVIDIA_GPU_MODEL(str, Enum):
+    UNKNOWN = "UNKNOWN"
+    A100_40GB = GPU_MODEL.A100_40GB
+    A100_80GB = GPU_MODEL.A100_80GB
+    H100_80GB = GPU_MODEL.H100
+
+
+class NvidiaGPUInfo(BaseModel):
+    index: str
+    uuid: str
+    product_name: str  # The official product name.
+    model: NVIDIA_GPU_MODEL
+
+
+def nvidia_gpus_are_available() -> bool:
+    try:
+        result: subprocess.CompletedProcess = subprocess.run(
+            ["nvidia-smi"], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL
+        )
+        return result.returncode == 0
+    except Exception:
+        return False
+
+
+def fetch_nvidia_gpu_infos(logger: Any) -> List[NvidiaGPUInfo]:
+    logger = logger.bind(module=__name__)
+    logger.info("Fetching GPU information")
+
+    try:
+        result: subprocess.CompletedProcess = subprocess.run(
+            ["nvidia-smi", "--query-gpu=index,name,uuid", "--format=csv,noheader"],
+            capture_output=True,
+            check=True,
+            text=True,
+        )
+    except subprocess.CalledProcessError as e:
+        logger.error("Failed to fetch GPU information", exc_info=e)
+        raise
+
+    infos: List[NvidiaGPUInfo] = []
+    for line in result.stdout.splitlines():
+        # Example:
+        # nvidia-smi --query-gpu=index,name,uuid --format=csv,noheader
+        # 0, NVIDIA A100-SXM4-80GB, GPU-89fdc1e1-18b2-f499-c12b-82bcb9bfb3fa
+        # 1, NVIDIA A100-PCIE-40GB, GPU-e9c9aa65-bff3-405a-ab7c-dc879cc88169
+        # 2, NVIDIA H100 80GB HBM3, GPU-8c35f4c9-4dff-c9a2-866f-afb5d82e1dd7
+        parts = line.split(",")
+        index = parts[0].strip()
+        product_name = parts[1].strip()
+        uuid = parts[2].strip()
+
+        model = NVIDIA_GPU_MODEL.UNKNOWN
+        if product_name.startswith("NVIDIA A100") and product_name.endswith("80GB"):
+            model = NVIDIA_GPU_MODEL.A100_80GB
+        if product_name.startswith("NVIDIA A100") and product_name.endswith("40GB"):
+            model = NVIDIA_GPU_MODEL.A100_40GB
+        elif product_name.startswith("NVIDIA H100"):
+            model = NVIDIA_GPU_MODEL.H100_80GB
+
+        if model == NVIDIA_GPU_MODEL.UNKNOWN:
+            logger.warning("Unknown GPU model detected", nvidia_smi_output=line)
+
+        infos.append(
+            NvidiaGPUInfo(
+                index=index, uuid=uuid, product_name=product_name, model=model
+            )
+        )
+    return infos

indexify/executor/host_resources/nvidia_gpu_allocator.py

@@ -0,0 +1,52 @@
+from typing import Any, List
+
+from .nvidia_gpu import NvidiaGPUInfo, fetch_nvidia_gpu_infos, nvidia_gpus_are_available
+
+
+class NvidiaGPUAllocator:
+    """NvidiaGPUAllocator is a class that manages the allocation and deallocation of GPUs."""
+
+    def __init__(self, logger: Any):
+        gpu_infos: List[NvidiaGPUInfo] = []
+
+        if nvidia_gpus_are_available():
+            gpu_infos = fetch_nvidia_gpu_infos(logger)
+            logger.bind(module=__name__).info(
+                "Fetched information about NVIDIA GPUs:", info=gpu_infos
+            )
+
+        self._all_gpus: List[NvidiaGPUInfo] = gpu_infos
+        self._free_gpus: List[NvidiaGPUInfo] = list(gpu_infos)
+
+    def allocate(self, count: int, logger: Any) -> List[NvidiaGPUInfo]:
+        """
+        Allocates a specified number of GPUs.
+
+        Args:
+            count (int): The number of GPUs to allocate.
+
+        Returns:
+            List[NvidiaGPUInfo]: A list of allocated GPUs. The list is empty if count is 0.
+
+        Raises:
+            ValueError: If the requested number of GPUs exceeds free GPUs.
+            Exception: If an error occurs during allocation.
+        """
+        if count > len(self._free_gpus):
+            raise ValueError(
+                f"Not enough free GPUs available, requested={count}, available={len(self._free_gpus)}"
+            )
+
+        allocated_gpus: List[NvidiaGPUInfo] = []
+        for _ in range(count):
+            allocated_gpus.append(self._free_gpus.pop())
+        return allocated_gpus
+
+    def deallocate(self, gpus: List[NvidiaGPUInfo]) -> None:
+        self._free_gpus.extend(gpus)
+
+    def list_all(self) -> List[NvidiaGPUInfo]:
+        return list(self._all_gpus)  # Return a copy to avoid external modification
+
+    def list_free(self) -> List[NvidiaGPUInfo]:
+        return list(self._free_gpus)  # Return a copy to avoid external modification

indexify/proto/executor_api.proto

@@ -28,26 +28,17 @@ message DataPayload {
 
 // ===== report_executor_state RPC =====
 
+// Deprecated enum. TODO: remove when all the code is using model string.
 enum GPUModel {
     GPU_MODEL_UNKNOWN = 0;
-    GPU_MODEL_NVIDIA_TESLA_T4_16GB = 10;
-    GPU_MODEL_NVIDIA_TESLA_V100_16GB = 20;
-    GPU_MODEL_NVIDIA_A10_24GB = 30;
-    GPU_MODEL_NVIDIA_A6000_48GB = 40;
-    // A100 GPUs
-    GPU_MODEL_NVIDIA_A100_SXM4_40GB = 50;
-    GPU_MODEL_NVIDIA_A100_SXM4_80GB = 51;
-    GPU_MODEL_NVIDIA_A100_PCI_40GB = 52;
-    // H100 GPUs
-    GPU_MODEL_NVIDIA_H100_SXM5_80GB = 60;
-    GPU_MODEL_NVIDIA_H100_PCI_80GB = 61;
-    GPU_MODEL_NVIDIA_RTX_6000_24GB = 62;
 }
 
 // Free GPUs available at the Executor.
 message GPUResources {
     optional uint32 count = 1;
-    optional GPUModel model = 2;
+    optional GPUModel deprecated_model = 2;
+    // Either GPU_MODEL value from Tensorlake SDK or "UNKNOWN"
+    optional string model = 3;
 }
 
 // Resources that we're currently tracking and limiting on Executor.
@@ -55,7 +46,8 @@ message HostResources {
     optional uint32 cpu_count = 1;
     optional uint64 memory_bytes = 2;
     optional uint64 disk_bytes = 3;
-    optional GPUResources gpu = 4;
+    // Not set if no GPUs are available.
+    optional GPUResources gpu = 4; 
 }
 
 // Specification of a single function that is allowed to be run on the Executor.
@@ -83,6 +75,15 @@ enum FunctionExecutorStatus {
 }
 
 // Immutable information that identifies and describes a Function Executor.
+message FunctionExecutorResources {
+    // 1000 CPU ms per sec is one full CPU core.
+    // 2000 CPU ms per sec is two full CPU cores.
+    optional uint32 cpu_ms_per_sec = 1;
+    optional uint64 memory_bytes = 2;
+    optional uint64 disk_bytes = 3;
+    optional uint32 gpu_count = 4;
+}
+
 message FunctionExecutorDescription {
     optional string id = 1;
     optional string namespace = 2;
@@ -95,6 +96,7 @@ message FunctionExecutorDescription {
     // Timeout for customer code duration during FE creation.
     optional uint32 customer_code_timeout_ms = 9;
     optional DataPayload graph = 10;
+    optional FunctionExecutorResources resources = 11;
 }
 
 message FunctionExecutorState {
@@ -128,8 +130,10 @@ message ExecutorState {
     optional ExecutorFlavor flavor = 4;
     optional string version = 5;
     optional ExecutorStatus status = 6;
+    // Total resources available at the Executor.
+    optional HostResources total_resources = 13;
     // Free resources available at the Executor.
-    optional HostResources free_resources = 7;
+    optional HostResources free_resources = 7; // Not used right now.
     // Empty allowed_functions list means that any function can run on the Executor.
     repeated AllowedFunction allowed_functions = 8;
     repeated FunctionExecutorState function_executor_states = 9;
@@ -150,6 +154,14 @@ message ReportExecutorStateResponse {
 }
 
 // ===== get_desired_executor_states RPC =====
+message TaskRetryPolicy {
+    optional uint32 max_retries = 1;
+    optional uint32 initial_delay_ms = 2;
+    optional uint32 max_delay_ms = 3;
+    // The multiplier value is 1000x of the actual value to avoid working with floating point.
+    optional uint32 delay_multiplier = 4;
+}
+
 message Task {
     optional string id = 1;
     optional string namespace = 2;
@@ -166,6 +178,7 @@ message Task {
     // S3 URI if the data is stored in S3.
     // Starts with "file://"" prefix followed by an absolute directory path if the data is stored on a local file system.
     optional string output_payload_uri_prefix = 13;
+    optional TaskRetryPolicy retry_policy = 14;
 }
 
 message TaskAllocation {

indexify/proto/executor_api_pb2.py

@@ -19,7 +19,7 @@ _sym_db = _symbol_database.Default()
 
 
 DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(
-    b'\n!indexify/proto/executor_api.proto\x12\x0f\x65xecutor_api_pb"\x87\x02\n\x0b\x44\x61taPayload\x12\x11\n\x04path\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x11\n\x04size\x18\x02 \x01(\x04H\x01\x88\x01\x01\x12\x18\n\x0bsha256_hash\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x10\n\x03uri\x18\x04 \x01(\tH\x03\x88\x01\x01\x12;\n\x08\x65ncoding\x18\x05 \x01(\x0e\x32$.executor_api_pb.DataPayloadEncodingH\x04\x88\x01\x01\x12\x1d\n\x10\x65ncoding_version\x18\x06 \x01(\x04H\x05\x88\x01\x01\x42\x07\n\x05_pathB\x07\n\x05_sizeB\x0e\n\x0c_sha256_hashB\x06\n\x04_uriB\x0b\n\t_encodingB\x13\n\x11_encoding_version"e\n\x0cGPUResources\x12\x12\n\x05\x63ount\x18\x01 \x01(\rH\x00\x88\x01\x01\x12-\n\x05model\x18\x02 \x01(\x0e\x32\x19.executor_api_pb.GPUModelH\x01\x88\x01\x01\x42\x08\n\x06_countB\x08\n\x06_model"\xc2\x01\n\rHostResources\x12\x16\n\tcpu_count\x18\x01 \x01(\rH\x00\x88\x01\x01\x12\x19\n\x0cmemory_bytes\x18\x02 \x01(\x04H\x01\x88\x01\x01\x12\x17\n\ndisk_bytes\x18\x03 \x01(\x04H\x02\x88\x01\x01\x12/\n\x03gpu\x18\x04 \x01(\x0b\x32\x1d.executor_api_pb.GPUResourcesH\x03\x88\x01\x01\x42\x0c\n\n_cpu_countB\x0f\n\r_memory_bytesB\r\n\x0b_disk_bytesB\x06\n\x04_gpu"\xbb\x01\n\x0f\x41llowedFunction\x12\x16\n\tnamespace\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x17\n\ngraph_name\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x42\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_function_nameB\x10\n\x0e_graph_version"\xed\x03\n\x1b\x46unctionExecutorDescription\x12\x0f\n\x02id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x05 \x01(\tH\x04\x88\x01\x01\x12\x16\n\timage_uri\x18\x06 \x01(\tH\x05\x88\x01\x01\x12\x14\n\x0csecret_names\x18\x07 \x03(\t\x12<\n\x0fresource_limits\x18\x08 \x01(\x0b\x32\x1e.executor_api_pb.HostResourcesH\x06\x88\x01\x01\x12%\n\x18\x63ustomer_code_timeout_ms\x18\t \x01(\rH\x07\x88\x01\x01\x12\x30\n\x05graph\x18\n \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\x08\x88\x01\x01\x42\x05\n\x03_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_graph_versionB\x10\n\x0e_function_nameB\x0c\n\n_image_uriB\x12\n\x10_resource_limitsB\x1b\n\x19_customer_code_timeout_msB\x08\n\x06_graph"\xe8\x01\n\x15\x46unctionExecutorState\x12\x46\n\x0b\x64\x65scription\x18\x01 \x01(\x0b\x32,.executor_api_pb.FunctionExecutorDescriptionH\x00\x88\x01\x01\x12<\n\x06status\x18\x02 \x01(\x0e\x32\'.executor_api_pb.FunctionExecutorStatusH\x01\x88\x01\x01\x12\x1b\n\x0estatus_message\x18\x03 \x01(\tH\x02\x88\x01\x01\x42\x0e\n\x0c_descriptionB\t\n\x07_statusB\x11\n\x0f_status_message"\xcb\x05\n\rExecutorState\x12\x18\n\x0b\x65xecutor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x1d\n\x10\x64\x65velopment_mode\x18\x02 \x01(\x08H\x01\x88\x01\x01\x12\x15\n\x08hostname\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x34\n\x06\x66lavor\x18\x04 \x01(\x0e\x32\x1f.executor_api_pb.ExecutorFlavorH\x03\x88\x01\x01\x12\x14\n\x07version\x18\x05 \x01(\tH\x04\x88\x01\x01\x12\x34\n\x06status\x18\x06 \x01(\x0e\x32\x1f.executor_api_pb.ExecutorStatusH\x05\x88\x01\x01\x12;\n\x0e\x66ree_resources\x18\x07 \x01(\x0b\x32\x1e.executor_api_pb.HostResourcesH\x06\x88\x01\x01\x12;\n\x11\x61llowed_functions\x18\x08 \x03(\x0b\x32 .executor_api_pb.AllowedFunction\x12H\n\x18\x66unction_executor_states\x18\t \x03(\x0b\x32&.executor_api_pb.FunctionExecutorState\x12:\n\x06labels\x18\n \x03(\x0b\x32*.executor_api_pb.ExecutorState.LabelsEntry\x12\x17\n\nstate_hash\x18\x0b \x01(\tH\x07\x88\x01\x01\x12\x19\n\x0cserver_clock\x18\x0c \x01(\x04H\x08\x88\x01\x01\x1a-\n\x0bLabelsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x42\x0e\n\x0c_executor_idB\x13\n\x11_development_modeB\x0b\n\t_hostnameB\t\n\x07_flavorB\n\n\x08_versionB\t\n\x07_statusB\x11\n\x0f_free_resourcesB\r\n\x0b_state_hashB\x0f\n\r_server_clock"l\n\x1aReportExecutorStateRequest\x12;\n\x0e\x65xecutor_state\x18\x01 \x01(\x0b\x32\x1e.executor_api_pb.ExecutorStateH\x00\x88\x01\x01\x42\x11\n\x0f_executor_state"\x1d\n\x1bReportExecutorStateResponse"\xd6\x04\n\x04Task\x12\x0f\n\x02id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x05 \x01(\tH\x04\x88\x01\x01\x12 \n\x13graph_invocation_id\x18\x06 \x01(\tH\x05\x88\x01\x01\x12\x16\n\tinput_key\x18\x08 \x01(\tH\x06\x88\x01\x01\x12\x1f\n\x12reducer_output_key\x18\t \x01(\tH\x07\x88\x01\x01\x12\x17\n\ntimeout_ms\x18\n \x01(\rH\x08\x88\x01\x01\x12\x30\n\x05input\x18\x0b \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\t\x88\x01\x01\x12\x38\n\rreducer_input\x18\x0c \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\n\x88\x01\x01\x12&\n\x19output_payload_uri_prefix\x18\r \x01(\tH\x0b\x88\x01\x01\x42\x05\n\x03_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_graph_versionB\x10\n\x0e_function_nameB\x16\n\x14_graph_invocation_idB\x0c\n\n_input_keyB\x15\n\x13_reducer_output_keyB\r\n\x0b_timeout_msB\x08\n\x06_inputB\x10\n\x0e_reducer_inputB\x1c\n\x1a_output_payload_uri_prefix"\x7f\n\x0eTaskAllocation\x12!\n\x14\x66unction_executor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12(\n\x04task\x18\x02 \x01(\x0b\x32\x15.executor_api_pb.TaskH\x01\x88\x01\x01\x42\x17\n\x15_function_executor_idB\x07\n\x05_task"K\n\x1fGetDesiredExecutorStatesRequest\x12\x18\n\x0b\x65xecutor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x42\x0e\n\x0c_executor_id"\xb9\x01\n\x14\x44\x65siredExecutorState\x12H\n\x12\x66unction_executors\x18\x01 \x03(\x0b\x32,.executor_api_pb.FunctionExecutorDescription\x12\x39\n\x10task_allocations\x18\x02 \x03(\x0b\x32\x1f.executor_api_pb.TaskAllocation\x12\x12\n\x05\x63lock\x18\x03 \x01(\x04H\x00\x88\x01\x01\x42\x08\n\x06_clock"\x87\x06\n\x18ReportTaskOutcomeRequest\x12\x14\n\x07task_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x04 \x01(\tH\x03\x88\x01\x01\x12 \n\x13graph_invocation_id\x18\x06 \x01(\tH\x04\x88\x01\x01\x12\x32\n\x07outcome\x18\x07 \x01(\x0e\x32\x1c.executor_api_pb.TaskOutcomeH\x05\x88\x01\x01\x12\x1a\n\rinvocation_id\x18\x08 \x01(\tH\x06\x88\x01\x01\x12\x18\n\x0b\x65xecutor_id\x18\t \x01(\tH\x07\x88\x01\x01\x12\x14\n\x07reducer\x18\n \x01(\x08H\x08\x88\x01\x01\x12\x16\n\x0enext_functions\x18\x0b \x03(\t\x12\x30\n\nfn_outputs\x18\x0c \x03(\x0b\x32\x1c.executor_api_pb.DataPayload\x12\x31\n\x06stdout\x18\x0e \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\t\x88\x01\x01\x12\x31\n\x06stderr\x18\x0f \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\n\x88\x01\x01\x12=\n\x0foutput_encoding\x18\r \x01(\x0e\x32\x1f.executor_api_pb.OutputEncodingH\x0b\x88\x01\x01\x12$\n\x17output_encoding_version\x18\x05 \x01(\x04H\x0c\x88\x01\x01\x42\n\n\x08_task_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_function_nameB\x16\n\x14_graph_invocation_idB\n\n\x08_outcomeB\x10\n\x0e_invocation_idB\x0e\n\x0c_executor_idB\n\n\x08_reducerB\t\n\x07_stdoutB\t\n\x07_stderrB\x12\n\x10_output_encodingB\x1a\n\x18_output_encoding_version"\x1b\n\x19ReportTaskOutcomeResponse*\xab\x01\n\x13\x44\x61taPayloadEncoding\x12!\n\x1d\x44\x41TA_PAYLOAD_ENCODING_UNKNOWN\x10\x00\x12#\n\x1f\x44\x41TA_PAYLOAD_ENCODING_UTF8_JSON\x10\x01\x12#\n\x1f\x44\x41TA_PAYLOAD_ENCODING_UTF8_TEXT\x10\x02\x12\'\n#DATA_PAYLOAD_ENCODING_BINARY_PICKLE\x10\x03*\x86\x03\n\x08GPUModel\x12\x15\n\x11GPU_MODEL_UNKNOWN\x10\x00\x12"\n\x1eGPU_MODEL_NVIDIA_TESLA_T4_16GB\x10\n\x12$\n GPU_MODEL_NVIDIA_TESLA_V100_16GB\x10\x14\x12\x1d\n\x19GPU_MODEL_NVIDIA_A10_24GB\x10\x1e\x12\x1f\n\x1bGPU_MODEL_NVIDIA_A6000_48GB\x10(\x12#\n\x1fGPU_MODEL_NVIDIA_A100_SXM4_40GB\x10\x32\x12#\n\x1fGPU_MODEL_NVIDIA_A100_SXM4_80GB\x10\x33\x12"\n\x1eGPU_MODEL_NVIDIA_A100_PCI_40GB\x10\x34\x12#\n\x1fGPU_MODEL_NVIDIA_H100_SXM5_80GB\x10<\x12"\n\x1eGPU_MODEL_NVIDIA_H100_PCI_80GB\x10=\x12"\n\x1eGPU_MODEL_NVIDIA_RTX_6000_24GB\x10>*\xca\x03\n\x16\x46unctionExecutorStatus\x12$\n FUNCTION_EXECUTOR_STATUS_UNKNOWN\x10\x00\x12(\n$FUNCTION_EXECUTOR_STATUS_STARTING_UP\x10\x01\x12:\n6FUNCTION_EXECUTOR_STATUS_STARTUP_FAILED_CUSTOMER_ERROR\x10\x02\x12:\n6FUNCTION_EXECUTOR_STATUS_STARTUP_FAILED_PLATFORM_ERROR\x10\x03\x12!\n\x1d\x46UNCTION_EXECUTOR_STATUS_IDLE\x10\x04\x12)\n%FUNCTION_EXECUTOR_STATUS_RUNNING_TASK\x10\x05\x12&\n"FUNCTION_EXECUTOR_STATUS_UNHEALTHY\x10\x06\x12%\n!FUNCTION_EXECUTOR_STATUS_STOPPING\x10\x07\x12$\n FUNCTION_EXECUTOR_STATUS_STOPPED\x10\x08\x12%\n!FUNCTION_EXECUTOR_STATUS_SHUTDOWN\x10\t*\xc3\x01\n\x0e\x45xecutorStatus\x12\x1b\n\x17\x45XECUTOR_STATUS_UNKNOWN\x10\x00\x12\x1f\n\x1b\x45XECUTOR_STATUS_STARTING_UP\x10\x01\x12\x1b\n\x17\x45XECUTOR_STATUS_RUNNING\x10\x02\x12\x1b\n\x17\x45XECUTOR_STATUS_DRAINED\x10\x03\x12\x1c\n\x18\x45XECUTOR_STATUS_STOPPING\x10\x04\x12\x1b\n\x17\x45XECUTOR_STATUS_STOPPED\x10\x05*d\n\x0e\x45xecutorFlavor\x12\x1b\n\x17\x45XECUTOR_FLAVOR_UNKNOWN\x10\x00\x12\x17\n\x13\x45XECUTOR_FLAVOR_OSS\x10\x01\x12\x1c\n\x18\x45XECUTOR_FLAVOR_PLATFORM\x10\x02*[\n\x0bTaskOutcome\x12\x18\n\x14TASK_OUTCOME_UNKNOWN\x10\x00\x12\x18\n\x14TASK_OUTCOME_SUCCESS\x10\x01\x12\x18\n\x14TASK_OUTCOME_FAILURE\x10\x02*\x7f\n\x0eOutputEncoding\x12\x1b\n\x17OUTPUT_ENCODING_UNKNOWN\x10\x00\x12\x18\n\x14OUTPUT_ENCODING_JSON\x10\x01\x12\x1a\n\x16OUTPUT_ENCODING_PICKLE\x10\x02\x12\x1a\n\x16OUTPUT_ENCODING_BINARY\x10\x03\x32\xef\x02\n\x0b\x45xecutorAPI\x12t\n\x15report_executor_state\x12+.executor_api_pb.ReportExecutorStateRequest\x1a,.executor_api_pb.ReportExecutorStateResponse"\x00\x12z\n\x1bget_desired_executor_states\x12\x30.executor_api_pb.GetDesiredExecutorStatesRequest\x1a%.executor_api_pb.DesiredExecutorState"\x00\x30\x01\x12n\n\x13report_task_outcome\x12).executor_api_pb.ReportTaskOutcomeRequest\x1a*.executor_api_pb.ReportTaskOutcomeResponse"\x00\x62\x06proto3'
+    b'\n!indexify/proto/executor_api.proto\x12\x0f\x65xecutor_api_pb"\x87\x02\n\x0b\x44\x61taPayload\x12\x11\n\x04path\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x11\n\x04size\x18\x02 \x01(\x04H\x01\x88\x01\x01\x12\x18\n\x0bsha256_hash\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x10\n\x03uri\x18\x04 \x01(\tH\x03\x88\x01\x01\x12;\n\x08\x65ncoding\x18\x05 \x01(\x0e\x32$.executor_api_pb.DataPayloadEncodingH\x04\x88\x01\x01\x12\x1d\n\x10\x65ncoding_version\x18\x06 \x01(\x04H\x05\x88\x01\x01\x42\x07\n\x05_pathB\x07\n\x05_sizeB\x0e\n\x0c_sha256_hashB\x06\n\x04_uriB\x0b\n\t_encodingB\x13\n\x11_encoding_version"\x99\x01\n\x0cGPUResources\x12\x12\n\x05\x63ount\x18\x01 \x01(\rH\x00\x88\x01\x01\x12\x38\n\x10\x64\x65precated_model\x18\x02 \x01(\x0e\x32\x19.executor_api_pb.GPUModelH\x01\x88\x01\x01\x12\x12\n\x05model\x18\x03 \x01(\tH\x02\x88\x01\x01\x42\x08\n\x06_countB\x13\n\x11_deprecated_modelB\x08\n\x06_model"\xc2\x01\n\rHostResources\x12\x16\n\tcpu_count\x18\x01 \x01(\rH\x00\x88\x01\x01\x12\x19\n\x0cmemory_bytes\x18\x02 \x01(\x04H\x01\x88\x01\x01\x12\x17\n\ndisk_bytes\x18\x03 \x01(\x04H\x02\x88\x01\x01\x12/\n\x03gpu\x18\x04 \x01(\x0b\x32\x1d.executor_api_pb.GPUResourcesH\x03\x88\x01\x01\x42\x0c\n\n_cpu_countB\x0f\n\r_memory_bytesB\r\n\x0b_disk_bytesB\x06\n\x04_gpu"\xbb\x01\n\x0f\x41llowedFunction\x12\x16\n\tnamespace\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x17\n\ngraph_name\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x42\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_function_nameB\x10\n\x0e_graph_version"\xc5\x01\n\x19\x46unctionExecutorResources\x12\x1b\n\x0e\x63pu_ms_per_sec\x18\x01 \x01(\rH\x00\x88\x01\x01\x12\x19\n\x0cmemory_bytes\x18\x02 \x01(\x04H\x01\x88\x01\x01\x12\x17\n\ndisk_bytes\x18\x03 \x01(\x04H\x02\x88\x01\x01\x12\x16\n\tgpu_count\x18\x04 \x01(\rH\x03\x88\x01\x01\x42\x11\n\x0f_cpu_ms_per_secB\x0f\n\r_memory_bytesB\r\n\x0b_disk_bytesB\x0c\n\n_gpu_count"\xbf\x04\n\x1b\x46unctionExecutorDescription\x12\x0f\n\x02id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x05 \x01(\tH\x04\x88\x01\x01\x12\x16\n\timage_uri\x18\x06 \x01(\tH\x05\x88\x01\x01\x12\x14\n\x0csecret_names\x18\x07 \x03(\t\x12<\n\x0fresource_limits\x18\x08 \x01(\x0b\x32\x1e.executor_api_pb.HostResourcesH\x06\x88\x01\x01\x12%\n\x18\x63ustomer_code_timeout_ms\x18\t \x01(\rH\x07\x88\x01\x01\x12\x30\n\x05graph\x18\n \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\x08\x88\x01\x01\x12\x42\n\tresources\x18\x0b \x01(\x0b\x32*.executor_api_pb.FunctionExecutorResourcesH\t\x88\x01\x01\x42\x05\n\x03_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_graph_versionB\x10\n\x0e_function_nameB\x0c\n\n_image_uriB\x12\n\x10_resource_limitsB\x1b\n\x19_customer_code_timeout_msB\x08\n\x06_graphB\x0c\n\n_resources"\xe8\x01\n\x15\x46unctionExecutorState\x12\x46\n\x0b\x64\x65scription\x18\x01 \x01(\x0b\x32,.executor_api_pb.FunctionExecutorDescriptionH\x00\x88\x01\x01\x12<\n\x06status\x18\x02 \x01(\x0e\x32\'.executor_api_pb.FunctionExecutorStatusH\x01\x88\x01\x01\x12\x1b\n\x0estatus_message\x18\x03 \x01(\tH\x02\x88\x01\x01\x42\x0e\n\x0c_descriptionB\t\n\x07_statusB\x11\n\x0f_status_message"\x9d\x06\n\rExecutorState\x12\x18\n\x0b\x65xecutor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x1d\n\x10\x64\x65velopment_mode\x18\x02 \x01(\x08H\x01\x88\x01\x01\x12\x15\n\x08hostname\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x34\n\x06\x66lavor\x18\x04 \x01(\x0e\x32\x1f.executor_api_pb.ExecutorFlavorH\x03\x88\x01\x01\x12\x14\n\x07version\x18\x05 \x01(\tH\x04\x88\x01\x01\x12\x34\n\x06status\x18\x06 \x01(\x0e\x32\x1f.executor_api_pb.ExecutorStatusH\x05\x88\x01\x01\x12<\n\x0ftotal_resources\x18\r \x01(\x0b\x32\x1e.executor_api_pb.HostResourcesH\x06\x88\x01\x01\x12;\n\x0e\x66ree_resources\x18\x07 \x01(\x0b\x32\x1e.executor_api_pb.HostResourcesH\x07\x88\x01\x01\x12;\n\x11\x61llowed_functions\x18\x08 \x03(\x0b\x32 .executor_api_pb.AllowedFunction\x12H\n\x18\x66unction_executor_states\x18\t \x03(\x0b\x32&.executor_api_pb.FunctionExecutorState\x12:\n\x06labels\x18\n \x03(\x0b\x32*.executor_api_pb.ExecutorState.LabelsEntry\x12\x17\n\nstate_hash\x18\x0b \x01(\tH\x08\x88\x01\x01\x12\x19\n\x0cserver_clock\x18\x0c \x01(\x04H\t\x88\x01\x01\x1a-\n\x0bLabelsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x42\x0e\n\x0c_executor_idB\x13\n\x11_development_modeB\x0b\n\t_hostnameB\t\n\x07_flavorB\n\n\x08_versionB\t\n\x07_statusB\x12\n\x10_total_resourcesB\x11\n\x0f_free_resourcesB\r\n\x0b_state_hashB\x0f\n\r_server_clock"l\n\x1aReportExecutorStateRequest\x12;\n\x0e\x65xecutor_state\x18\x01 \x01(\x0b\x32\x1e.executor_api_pb.ExecutorStateH\x00\x88\x01\x01\x42\x11\n\x0f_executor_state"\x1d\n\x1bReportExecutorStateResponse"\xcf\x01\n\x0fTaskRetryPolicy\x12\x18\n\x0bmax_retries\x18\x01 \x01(\rH\x00\x88\x01\x01\x12\x1d\n\x10initial_delay_ms\x18\x02 \x01(\rH\x01\x88\x01\x01\x12\x19\n\x0cmax_delay_ms\x18\x03 \x01(\rH\x02\x88\x01\x01\x12\x1d\n\x10\x64\x65lay_multiplier\x18\x04 \x01(\rH\x03\x88\x01\x01\x42\x0e\n\x0c_max_retriesB\x13\n\x11_initial_delay_msB\x0f\n\r_max_delay_msB\x13\n\x11_delay_multiplier"\xa4\x05\n\x04Task\x12\x0f\n\x02id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rgraph_version\x18\x04 \x01(\tH\x03\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x05 \x01(\tH\x04\x88\x01\x01\x12 \n\x13graph_invocation_id\x18\x06 \x01(\tH\x05\x88\x01\x01\x12\x16\n\tinput_key\x18\x08 \x01(\tH\x06\x88\x01\x01\x12\x1f\n\x12reducer_output_key\x18\t \x01(\tH\x07\x88\x01\x01\x12\x17\n\ntimeout_ms\x18\n \x01(\rH\x08\x88\x01\x01\x12\x30\n\x05input\x18\x0b \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\t\x88\x01\x01\x12\x38\n\rreducer_input\x18\x0c \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\n\x88\x01\x01\x12&\n\x19output_payload_uri_prefix\x18\r \x01(\tH\x0b\x88\x01\x01\x12;\n\x0cretry_policy\x18\x0e \x01(\x0b\x32 .executor_api_pb.TaskRetryPolicyH\x0c\x88\x01\x01\x42\x05\n\x03_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_graph_versionB\x10\n\x0e_function_nameB\x16\n\x14_graph_invocation_idB\x0c\n\n_input_keyB\x15\n\x13_reducer_output_keyB\r\n\x0b_timeout_msB\x08\n\x06_inputB\x10\n\x0e_reducer_inputB\x1c\n\x1a_output_payload_uri_prefixB\x0f\n\r_retry_policy"\x7f\n\x0eTaskAllocation\x12!\n\x14\x66unction_executor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12(\n\x04task\x18\x02 \x01(\x0b\x32\x15.executor_api_pb.TaskH\x01\x88\x01\x01\x42\x17\n\x15_function_executor_idB\x07\n\x05_task"K\n\x1fGetDesiredExecutorStatesRequest\x12\x18\n\x0b\x65xecutor_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x42\x0e\n\x0c_executor_id"\xb9\x01\n\x14\x44\x65siredExecutorState\x12H\n\x12\x66unction_executors\x18\x01 \x03(\x0b\x32,.executor_api_pb.FunctionExecutorDescription\x12\x39\n\x10task_allocations\x18\x02 \x03(\x0b\x32\x1f.executor_api_pb.TaskAllocation\x12\x12\n\x05\x63lock\x18\x03 \x01(\x04H\x00\x88\x01\x01\x42\x08\n\x06_clock"\x87\x06\n\x18ReportTaskOutcomeRequest\x12\x14\n\x07task_id\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x16\n\tnamespace\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x17\n\ngraph_name\x18\x03 \x01(\tH\x02\x88\x01\x01\x12\x1a\n\rfunction_name\x18\x04 \x01(\tH\x03\x88\x01\x01\x12 \n\x13graph_invocation_id\x18\x06 \x01(\tH\x04\x88\x01\x01\x12\x32\n\x07outcome\x18\x07 \x01(\x0e\x32\x1c.executor_api_pb.TaskOutcomeH\x05\x88\x01\x01\x12\x1a\n\rinvocation_id\x18\x08 \x01(\tH\x06\x88\x01\x01\x12\x18\n\x0b\x65xecutor_id\x18\t \x01(\tH\x07\x88\x01\x01\x12\x14\n\x07reducer\x18\n \x01(\x08H\x08\x88\x01\x01\x12\x16\n\x0enext_functions\x18\x0b \x03(\t\x12\x30\n\nfn_outputs\x18\x0c \x03(\x0b\x32\x1c.executor_api_pb.DataPayload\x12\x31\n\x06stdout\x18\x0e \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\t\x88\x01\x01\x12\x31\n\x06stderr\x18\x0f \x01(\x0b\x32\x1c.executor_api_pb.DataPayloadH\n\x88\x01\x01\x12=\n\x0foutput_encoding\x18\r \x01(\x0e\x32\x1f.executor_api_pb.OutputEncodingH\x0b\x88\x01\x01\x12$\n\x17output_encoding_version\x18\x05 \x01(\x04H\x0c\x88\x01\x01\x42\n\n\x08_task_idB\x0c\n\n_namespaceB\r\n\x0b_graph_nameB\x10\n\x0e_function_nameB\x16\n\x14_graph_invocation_idB\n\n\x08_outcomeB\x10\n\x0e_invocation_idB\x0e\n\x0c_executor_idB\n\n\x08_reducerB\t\n\x07_stdoutB\t\n\x07_stderrB\x12\n\x10_output_encodingB\x1a\n\x18_output_encoding_version"\x1b\n\x19ReportTaskOutcomeResponse*\xab\x01\n\x13\x44\x61taPayloadEncoding\x12!\n\x1d\x44\x41TA_PAYLOAD_ENCODING_UNKNOWN\x10\x00\x12#\n\x1f\x44\x41TA_PAYLOAD_ENCODING_UTF8_JSON\x10\x01\x12#\n\x1f\x44\x41TA_PAYLOAD_ENCODING_UTF8_TEXT\x10\x02\x12\'\n#DATA_PAYLOAD_ENCODING_BINARY_PICKLE\x10\x03*!\n\x08GPUModel\x12\x15\n\x11GPU_MODEL_UNKNOWN\x10\x00*\xca\x03\n\x16\x46unctionExecutorStatus\x12$\n FUNCTION_EXECUTOR_STATUS_UNKNOWN\x10\x00\x12(\n$FUNCTION_EXECUTOR_STATUS_STARTING_UP\x10\x01\x12:\n6FUNCTION_EXECUTOR_STATUS_STARTUP_FAILED_CUSTOMER_ERROR\x10\x02\x12:\n6FUNCTION_EXECUTOR_STATUS_STARTUP_FAILED_PLATFORM_ERROR\x10\x03\x12!\n\x1d\x46UNCTION_EXECUTOR_STATUS_IDLE\x10\x04\x12)\n%FUNCTION_EXECUTOR_STATUS_RUNNING_TASK\x10\x05\x12&\n"FUNCTION_EXECUTOR_STATUS_UNHEALTHY\x10\x06\x12%\n!FUNCTION_EXECUTOR_STATUS_STOPPING\x10\x07\x12$\n FUNCTION_EXECUTOR_STATUS_STOPPED\x10\x08\x12%\n!FUNCTION_EXECUTOR_STATUS_SHUTDOWN\x10\t*\xc3\x01\n\x0e\x45xecutorStatus\x12\x1b\n\x17\x45XECUTOR_STATUS_UNKNOWN\x10\x00\x12\x1f\n\x1b\x45XECUTOR_STATUS_STARTING_UP\x10\x01\x12\x1b\n\x17\x45XECUTOR_STATUS_RUNNING\x10\x02\x12\x1b\n\x17\x45XECUTOR_STATUS_DRAINED\x10\x03\x12\x1c\n\x18\x45XECUTOR_STATUS_STOPPING\x10\x04\x12\x1b\n\x17\x45XECUTOR_STATUS_STOPPED\x10\x05*d\n\x0e\x45xecutorFlavor\x12\x1b\n\x17\x45XECUTOR_FLAVOR_UNKNOWN\x10\x00\x12\x17\n\x13\x45XECUTOR_FLAVOR_OSS\x10\x01\x12\x1c\n\x18\x45XECUTOR_FLAVOR_PLATFORM\x10\x02*[\n\x0bTaskOutcome\x12\x18\n\x14TASK_OUTCOME_UNKNOWN\x10\x00\x12\x18\n\x14TASK_OUTCOME_SUCCESS\x10\x01\x12\x18\n\x14TASK_OUTCOME_FAILURE\x10\x02*\x7f\n\x0eOutputEncoding\x12\x1b\n\x17OUTPUT_ENCODING_UNKNOWN\x10\x00\x12\x18\n\x14OUTPUT_ENCODING_JSON\x10\x01\x12\x1a\n\x16OUTPUT_ENCODING_PICKLE\x10\x02\x12\x1a\n\x16OUTPUT_ENCODING_BINARY\x10\x03\x32\xef\x02\n\x0b\x45xecutorAPI\x12t\n\x15report_executor_state\x12+.executor_api_pb.ReportExecutorStateRequest\x1a,.executor_api_pb.ReportExecutorStateResponse"\x00\x12z\n\x1bget_desired_executor_states\x12\x30.executor_api_pb.GetDesiredExecutorStatesRequest\x1a%.executor_api_pb.DesiredExecutorState"\x00\x30\x01\x12n\n\x13report_task_outcome\x12).executor_api_pb.ReportTaskOutcomeRequest\x1a*.executor_api_pb.ReportTaskOutcomeResponse"\x00\x62\x06proto3'
 )
 
 _globals = globals()
@@ -31,52 +31,56 @@ if not _descriptor._USE_C_DESCRIPTORS:
     DESCRIPTOR._loaded_options = None
     _globals["_EXECUTORSTATE_LABELSENTRY"]._loaded_options = None
     _globals["_EXECUTORSTATE_LABELSENTRY"]._serialized_options = b"8\001"
-    _globals["_DATAPAYLOADENCODING"]._serialized_start = 4203
-    _globals["_DATAPAYLOADENCODING"]._serialized_end = 4374
-    _globals["_GPUMODEL"]._serialized_start = 4377
-    _globals["_GPUMODEL"]._serialized_end = 4767
-    _globals["_FUNCTIONEXECUTORSTATUS"]._serialized_start = 4770
-    _globals["_FUNCTIONEXECUTORSTATUS"]._serialized_end = 5228
-    _globals["_EXECUTORSTATUS"]._serialized_start = 5231
-    _globals["_EXECUTORSTATUS"]._serialized_end = 5426
-    _globals["_EXECUTORFLAVOR"]._serialized_start = 5428
-    _globals["_EXECUTORFLAVOR"]._serialized_end = 5528
-    _globals["_TASKOUTCOME"]._serialized_start = 5530
-    _globals["_TASKOUTCOME"]._serialized_end = 5621
-    _globals["_OUTPUTENCODING"]._serialized_start = 5623
-    _globals["_OUTPUTENCODING"]._serialized_end = 5750
+    _globals["_DATAPAYLOADENCODING"]._serialized_start = 4908
+    _globals["_DATAPAYLOADENCODING"]._serialized_end = 5079
+    _globals["_GPUMODEL"]._serialized_start = 5081
+    _globals["_GPUMODEL"]._serialized_end = 5114
+    _globals["_FUNCTIONEXECUTORSTATUS"]._serialized_start = 5117
+    _globals["_FUNCTIONEXECUTORSTATUS"]._serialized_end = 5575
+    _globals["_EXECUTORSTATUS"]._serialized_start = 5578
+    _globals["_EXECUTORSTATUS"]._serialized_end = 5773
+    _globals["_EXECUTORFLAVOR"]._serialized_start = 5775
+    _globals["_EXECUTORFLAVOR"]._serialized_end = 5875
+    _globals["_TASKOUTCOME"]._serialized_start = 5877
+    _globals["_TASKOUTCOME"]._serialized_end = 5968
+    _globals["_OUTPUTENCODING"]._serialized_start = 5970
+    _globals["_OUTPUTENCODING"]._serialized_end = 6097
     _globals["_DATAPAYLOAD"]._serialized_start = 55
     _globals["_DATAPAYLOAD"]._serialized_end = 318
-    _globals["_GPURESOURCES"]._serialized_start = 320
-    _globals["_GPURESOURCES"]._serialized_end = 421
-    _globals["_HOSTRESOURCES"]._serialized_start = 424
-    _globals["_HOSTRESOURCES"]._serialized_end = 618
-    _globals["_ALLOWEDFUNCTION"]._serialized_start = 621
-    _globals["_ALLOWEDFUNCTION"]._serialized_end = 808
-    _globals["_FUNCTIONEXECUTORDESCRIPTION"]._serialized_start = 811
-    _globals["_FUNCTIONEXECUTORDESCRIPTION"]._serialized_end = 1304
-    _globals["_FUNCTIONEXECUTORSTATE"]._serialized_start = 1307
-    _globals["_FUNCTIONEXECUTORSTATE"]._serialized_end = 1539
-    _globals["_EXECUTORSTATE"]._serialized_start = 1542
-    _globals["_EXECUTORSTATE"]._serialized_end = 2257
-    _globals["_EXECUTORSTATE_LABELSENTRY"]._serialized_start = 2077
-    _globals["_EXECUTORSTATE_LABELSENTRY"]._serialized_end = 2122
-    _globals["_REPORTEXECUTORSTATEREQUEST"]._serialized_start = 2259
-    _globals["_REPORTEXECUTORSTATEREQUEST"]._serialized_end = 2367
-    _globals["_REPORTEXECUTORSTATERESPONSE"]._serialized_start = 2369
-    _globals["_REPORTEXECUTORSTATERESPONSE"]._serialized_end = 2398
-    _globals["_TASK"]._serialized_start = 2401
-    _globals["_TASK"]._serialized_end = 2999
-    _globals["_TASKALLOCATION"]._serialized_start = 3001
-    _globals["_TASKALLOCATION"]._serialized_end = 3128
-    _globals["_GETDESIREDEXECUTORSTATESREQUEST"]._serialized_start = 3130
-    _globals["_GETDESIREDEXECUTORSTATESREQUEST"]._serialized_end = 3205
-    _globals["_DESIREDEXECUTORSTATE"]._serialized_start = 3208
-    _globals["_DESIREDEXECUTORSTATE"]._serialized_end = 3393
-    _globals["_REPORTTASKOUTCOMEREQUEST"]._serialized_start = 3396
-    _globals["_REPORTTASKOUTCOMEREQUEST"]._serialized_end = 4171
-    _globals["_REPORTTASKOUTCOMERESPONSE"]._serialized_start = 4173
-    _globals["_REPORTTASKOUTCOMERESPONSE"]._serialized_end = 4200
-    _globals["_EXECUTORAPI"]._serialized_start = 5753
-    _globals["_EXECUTORAPI"]._serialized_end = 6120
+    _globals["_GPURESOURCES"]._serialized_start = 321
+    _globals["_GPURESOURCES"]._serialized_end = 474
+    _globals["_HOSTRESOURCES"]._serialized_start = 477
+    _globals["_HOSTRESOURCES"]._serialized_end = 671
+    _globals["_ALLOWEDFUNCTION"]._serialized_start = 674
+    _globals["_ALLOWEDFUNCTION"]._serialized_end = 861
+    _globals["_FUNCTIONEXECUTORRESOURCES"]._serialized_start = 864
+    _globals["_FUNCTIONEXECUTORRESOURCES"]._serialized_end = 1061
+    _globals["_FUNCTIONEXECUTORDESCRIPTION"]._serialized_start = 1064
+    _globals["_FUNCTIONEXECUTORDESCRIPTION"]._serialized_end = 1639
+    _globals["_FUNCTIONEXECUTORSTATE"]._serialized_start = 1642
+    _globals["_FUNCTIONEXECUTORSTATE"]._serialized_end = 1874
+    _globals["_EXECUTORSTATE"]._serialized_start = 1877
+    _globals["_EXECUTORSTATE"]._serialized_end = 2674
+    _globals["_EXECUTORSTATE_LABELSENTRY"]._serialized_start = 2474
+    _globals["_EXECUTORSTATE_LABELSENTRY"]._serialized_end = 2519
+    _globals["_REPORTEXECUTORSTATEREQUEST"]._serialized_start = 2676
+    _globals["_REPORTEXECUTORSTATEREQUEST"]._serialized_end = 2784
+    _globals["_REPORTEXECUTORSTATERESPONSE"]._serialized_start = 2786
+    _globals["_REPORTEXECUTORSTATERESPONSE"]._serialized_end = 2815
+    _globals["_TASKRETRYPOLICY"]._serialized_start = 2818
+    _globals["_TASKRETRYPOLICY"]._serialized_end = 3025
+    _globals["_TASK"]._serialized_start = 3028
+    _globals["_TASK"]._serialized_end = 3704
+    _globals["_TASKALLOCATION"]._serialized_start = 3706
+    _globals["_TASKALLOCATION"]._serialized_end = 3833
+    _globals["_GETDESIREDEXECUTORSTATESREQUEST"]._serialized_start = 3835
+    _globals["_GETDESIREDEXECUTORSTATESREQUEST"]._serialized_end = 3910
+    _globals["_DESIREDEXECUTORSTATE"]._serialized_start = 3913
+    _globals["_DESIREDEXECUTORSTATE"]._serialized_end = 4098
+    _globals["_REPORTTASKOUTCOMEREQUEST"]._serialized_start = 4101
+    _globals["_REPORTTASKOUTCOMEREQUEST"]._serialized_end = 4876
+    _globals["_REPORTTASKOUTCOMERESPONSE"]._serialized_start = 4878
+    _globals["_REPORTTASKOUTCOMERESPONSE"]._serialized_end = 4905
+    _globals["_EXECUTORAPI"]._serialized_start = 6100
+    _globals["_EXECUTORAPI"]._serialized_end = 6467
 # @@protoc_insertion_point(module_scope)

indexify/proto/executor_api_pb2.pyi

@@ -21,16 +21,6 @@ class DataPayloadEncoding(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
 class GPUModel(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
     __slots__ = ()
     GPU_MODEL_UNKNOWN: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_TESLA_T4_16GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_TESLA_V100_16GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_A10_24GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_A6000_48GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_A100_SXM4_40GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_A100_SXM4_80GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_A100_PCI_40GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_H100_SXM5_80GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_H100_PCI_80GB: _ClassVar[GPUModel]
-    GPU_MODEL_NVIDIA_RTX_6000_24GB: _ClassVar[GPUModel]
 
 class FunctionExecutorStatus(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
     __slots__ = ()
@@ -82,16 +72,6 @@ DATA_PAYLOAD_ENCODING_UTF8_JSON: DataPayloadEncoding
 DATA_PAYLOAD_ENCODING_UTF8_TEXT: DataPayloadEncoding
 DATA_PAYLOAD_ENCODING_BINARY_PICKLE: DataPayloadEncoding
 GPU_MODEL_UNKNOWN: GPUModel
-GPU_MODEL_NVIDIA_TESLA_T4_16GB: GPUModel
-GPU_MODEL_NVIDIA_TESLA_V100_16GB: GPUModel
-GPU_MODEL_NVIDIA_A10_24GB: GPUModel
-GPU_MODEL_NVIDIA_A6000_48GB: GPUModel
-GPU_MODEL_NVIDIA_A100_SXM4_40GB: GPUModel
-GPU_MODEL_NVIDIA_A100_SXM4_80GB: GPUModel
-GPU_MODEL_NVIDIA_A100_PCI_40GB: GPUModel
-GPU_MODEL_NVIDIA_H100_SXM5_80GB: GPUModel
-GPU_MODEL_NVIDIA_H100_PCI_80GB: GPUModel
-GPU_MODEL_NVIDIA_RTX_6000_24GB: GPUModel
 FUNCTION_EXECUTOR_STATUS_UNKNOWN: FunctionExecutorStatus
 FUNCTION_EXECUTOR_STATUS_STARTING_UP: FunctionExecutorStatus
 FUNCTION_EXECUTOR_STATUS_STARTUP_FAILED_CUSTOMER_ERROR: FunctionExecutorStatus
@@ -144,13 +124,18 @@ class DataPayload(_message.Message):
     ) -> None: ...
 
 class GPUResources(_message.Message):
-    __slots__ = ("count", "model")
+    __slots__ = ("count", "deprecated_model", "model")
     COUNT_FIELD_NUMBER: _ClassVar[int]
+    DEPRECATED_MODEL_FIELD_NUMBER: _ClassVar[int]
     MODEL_FIELD_NUMBER: _ClassVar[int]
     count: int
-    model: GPUModel
+    deprecated_model: GPUModel
+    model: str
     def __init__(
-        self, count: _Optional[int] = ..., model: _Optional[_Union[GPUModel, str]] = ...
+        self,
+        count: _Optional[int] = ...,
+        deprecated_model: _Optional[_Union[GPUModel, str]] = ...,
+        model: _Optional[str] = ...,
     ) -> None: ...
 
 class HostResources(_message.Message):
@@ -189,6 +174,24 @@ class AllowedFunction(_message.Message):
         graph_version: _Optional[str] = ...,
     ) -> None: ...
 
+class FunctionExecutorResources(_message.Message):
+    __slots__ = ("cpu_ms_per_sec", "memory_bytes", "disk_bytes", "gpu_count")
+    CPU_MS_PER_SEC_FIELD_NUMBER: _ClassVar[int]
+    MEMORY_BYTES_FIELD_NUMBER: _ClassVar[int]
+    DISK_BYTES_FIELD_NUMBER: _ClassVar[int]
+    GPU_COUNT_FIELD_NUMBER: _ClassVar[int]
+    cpu_ms_per_sec: int
+    memory_bytes: int
+    disk_bytes: int
+    gpu_count: int
+    def __init__(
+        self,
+        cpu_ms_per_sec: _Optional[int] = ...,
+        memory_bytes: _Optional[int] = ...,
+        disk_bytes: _Optional[int] = ...,
+        gpu_count: _Optional[int] = ...,
+    ) -> None: ...
+
 class FunctionExecutorDescription(_message.Message):
     __slots__ = (
         "id",
@@ -201,6 +204,7 @@ class FunctionExecutorDescription(_message.Message):
         "resource_limits",
         "customer_code_timeout_ms",
         "graph",
+        "resources",
     )
     ID_FIELD_NUMBER: _ClassVar[int]
     NAMESPACE_FIELD_NUMBER: _ClassVar[int]
@@ -212,6 +216,7 @@ class FunctionExecutorDescription(_message.Message):
     RESOURCE_LIMITS_FIELD_NUMBER: _ClassVar[int]
     CUSTOMER_CODE_TIMEOUT_MS_FIELD_NUMBER: _ClassVar[int]
     GRAPH_FIELD_NUMBER: _ClassVar[int]
+    RESOURCES_FIELD_NUMBER: _ClassVar[int]
     id: str
     namespace: str
     graph_name: str
@@ -222,6 +227,7 @@ class FunctionExecutorDescription(_message.Message):
     resource_limits: HostResources
     customer_code_timeout_ms: int
     graph: DataPayload
+    resources: FunctionExecutorResources
     def __init__(
         self,
         id: _Optional[str] = ...,
@@ -234,6 +240,7 @@ class FunctionExecutorDescription(_message.Message):
         resource_limits: _Optional[_Union[HostResources, _Mapping]] = ...,
         customer_code_timeout_ms: _Optional[int] = ...,
         graph: _Optional[_Union[DataPayload, _Mapping]] = ...,
+        resources: _Optional[_Union[FunctionExecutorResources, _Mapping]] = ...,
     ) -> None: ...
 
 class FunctionExecutorState(_message.Message):
@@ -259,6 +266,7 @@ class ExecutorState(_message.Message):
         "flavor",
         "version",
         "status",
+        "total_resources",
         "free_resources",
         "allowed_functions",
         "function_executor_states",
@@ -283,6 +291,7 @@ class ExecutorState(_message.Message):
     FLAVOR_FIELD_NUMBER: _ClassVar[int]
     VERSION_FIELD_NUMBER: _ClassVar[int]
     STATUS_FIELD_NUMBER: _ClassVar[int]
+    TOTAL_RESOURCES_FIELD_NUMBER: _ClassVar[int]
     FREE_RESOURCES_FIELD_NUMBER: _ClassVar[int]
     ALLOWED_FUNCTIONS_FIELD_NUMBER: _ClassVar[int]
     FUNCTION_EXECUTOR_STATES_FIELD_NUMBER: _ClassVar[int]
@@ -295,6 +304,7 @@ class ExecutorState(_message.Message):
     flavor: ExecutorFlavor
     version: str
     status: ExecutorStatus
+    total_resources: HostResources
     free_resources: HostResources
     allowed_functions: _containers.RepeatedCompositeFieldContainer[AllowedFunction]
     function_executor_states: _containers.RepeatedCompositeFieldContainer[
@@ -311,6 +321,7 @@ class ExecutorState(_message.Message):
         flavor: _Optional[_Union[ExecutorFlavor, str]] = ...,
         version: _Optional[str] = ...,
         status: _Optional[_Union[ExecutorStatus, str]] = ...,
+        total_resources: _Optional[_Union[HostResources, _Mapping]] = ...,
         free_resources: _Optional[_Union[HostResources, _Mapping]] = ...,
         allowed_functions: _Optional[
             _Iterable[_Union[AllowedFunction, _Mapping]]
@@ -335,6 +346,24 @@ class ReportExecutorStateResponse(_message.Message):
     __slots__ = ()
     def __init__(self) -> None: ...
 
+class TaskRetryPolicy(_message.Message):
+    __slots__ = ("max_retries", "initial_delay_ms", "max_delay_ms", "delay_multiplier")
+    MAX_RETRIES_FIELD_NUMBER: _ClassVar[int]
+    INITIAL_DELAY_MS_FIELD_NUMBER: _ClassVar[int]
+    MAX_DELAY_MS_FIELD_NUMBER: _ClassVar[int]
+    DELAY_MULTIPLIER_FIELD_NUMBER: _ClassVar[int]
+    max_retries: int
+    initial_delay_ms: int
+    max_delay_ms: int
+    delay_multiplier: int
+    def __init__(
+        self,
+        max_retries: _Optional[int] = ...,
+        initial_delay_ms: _Optional[int] = ...,
+        max_delay_ms: _Optional[int] = ...,
+        delay_multiplier: _Optional[int] = ...,
+    ) -> None: ...
+
 class Task(_message.Message):
     __slots__ = (
         "id",
@@ -349,6 +378,7 @@ class Task(_message.Message):
         "input",
         "reducer_input",
         "output_payload_uri_prefix",
+        "retry_policy",
     )
     ID_FIELD_NUMBER: _ClassVar[int]
     NAMESPACE_FIELD_NUMBER: _ClassVar[int]
@@ -362,6 +392,7 @@ class Task(_message.Message):
     INPUT_FIELD_NUMBER: _ClassVar[int]
     REDUCER_INPUT_FIELD_NUMBER: _ClassVar[int]
     OUTPUT_PAYLOAD_URI_PREFIX_FIELD_NUMBER: _ClassVar[int]
+    RETRY_POLICY_FIELD_NUMBER: _ClassVar[int]
     id: str
     namespace: str
     graph_name: str
@@ -374,6 +405,7 @@ class Task(_message.Message):
     input: DataPayload
     reducer_input: DataPayload
     output_payload_uri_prefix: str
+    retry_policy: TaskRetryPolicy
     def __init__(
         self,
         id: _Optional[str] = ...,
@@ -388,6 +420,7 @@ class Task(_message.Message):
         input: _Optional[_Union[DataPayload, _Mapping]] = ...,
         reducer_input: _Optional[_Union[DataPayload, _Mapping]] = ...,
         output_payload_uri_prefix: _Optional[str] = ...,
+        retry_policy: _Optional[_Union[TaskRetryPolicy, _Mapping]] = ...,
     ) -> None: ...
 
 class TaskAllocation(_message.Message):

{indexify-0.3.22.dist-info → indexify-0.3.24.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: indexify
-Version: 0.3.22
+Version: 0.3.24
 Summary: Open Source Indexify components and helper tools
 Home-page: https://github.com/tensorlakeai/indexify
 License: Apache 2.0

{indexify-0.3.22.dist-info → indexify-0.3.24.dist-info}/RECORD

@@ -1,12 +1,12 @@
-indexify/cli/cli.py,sha256=k6vhgVKsHZvum2A2w7kbirpKusfqi_PyZGq29Q433_c,8950
+indexify/cli/cli.py,sha256=RaKrcdrawu_1FrDV67o2QwMqfsPdADOcIxK2iTZvcho,9190
 indexify/executor/README.md,sha256=ozC6_hMkhQQNVCMEpBxwiUALz6lwErPQxNxQfQDqnG4,2029
-indexify/executor/api_objects.py,sha256=UEMpD_kKr6BNfu2JdYzq1CPNrEmVZYhwlZRKf3hB7KI,1750
+indexify/executor/api_objects.py,sha256=kHx5gKPwM0Rm64Ea__kPFwuarStX0u_9uaE7vV5M5z8,2222
 indexify/executor/blob_store/blob_store.py,sha256=XViw_KRfFSNqwcFYwMZixZF-EYCjXK2AQHdt0xh4UVo,2368
 indexify/executor/blob_store/local_fs_blob_store.py,sha256=6LexqMBGXp8f6Ka95R6xMIUyDutrZJABOMNcp-ssa98,1809
 indexify/executor/blob_store/metrics/blob_store.py,sha256=5_xiPREeHWFtxFh1NupDsF8zP4pmUPgLNNn-UE9Uzvc,1008
 indexify/executor/blob_store/s3_blob_store.py,sha256=G3B_V3gUE7XbUY42lDtBczUKuA7q8S7MD43tx1aHrJo,3445
 indexify/executor/downloader.py,sha256=k9VbfOa-D6YH-cX8Sz-W-gWTsxmeVpSaIOq0xTC9KB0,15474
-indexify/executor/executor.py,sha256=K_xNJDImvIxgpE_ypw1ERyrijqOSe0EZIyVRVCqztVw,16697
+indexify/executor/executor.py,sha256=WE9ABct1yAlfh4-cOUcp1vTjFbkiXNAGlsbsNbfWdkU,17006
 indexify/executor/executor_flavor.py,sha256=uilzDQVVYlQGR1MVnrUC4NevUActDWHdnJkr38M6kTk,118
 indexify/executor/function_executor/function_executor.py,sha256=agfUxzSQ-2TqkpMhW3OvOSMF_EhpemetaL3_dYp29Ro,11888
 indexify/executor/function_executor/function_executor_state.py,sha256=ljPm1IrRMJ8hFklwvFp7Xax2HMpUIOHm0DwOxxMcy7U,4336
@@ -22,21 +22,24 @@ indexify/executor/function_executor/metrics/invocation_state_client.py,sha256=6F
 indexify/executor/function_executor/metrics/single_task_runner.py,sha256=7BJlGkdPGKeufMs3zWNO_1GRVzjINRY5rW3Mp4oWWec,805
 indexify/executor/function_executor/server/client_configuration.py,sha256=gOywMus0cotlX6NKIadEJwvOmBE-LbGE_wvoMi5-HzY,994
 indexify/executor/function_executor/server/function_executor_server.py,sha256=_DLivLDikupZusRk8gVWDk7fWPT9XjZ4un1yWSlOObs,883
-indexify/executor/function_executor/server/function_executor_server_factory.py,sha256=z13MGGCWU_G5SKUJpX4Qb7m5fYhIAH_-pqmTBh0WBQ8,1758
+indexify/executor/function_executor/server/function_executor_server_factory.py,sha256=xWEuDoxFqF-oC4RoiEer4S0Tk2tNbbJfA2kANZVShpM,1873
 indexify/executor/function_executor/server/subprocess_function_executor_server.py,sha256=JekDOqF7oFD4J6zcN3xB0Dxd1cgpEXMOsb_rKZOeBlI,668
 indexify/executor/function_executor/server/subprocess_function_executor_server_factory.py,sha256=g1AUbhOoPsdhp_50Ayahdyv1Ix5-nEBE8orOQfkATpM,4470
-indexify/executor/function_executor/single_task_runner.py,sha256=CWC2U1SilX6yVl1890quqSLYLexjRoBI-3LxNDEOISA,14335
+indexify/executor/function_executor/single_task_runner.py,sha256=6Fb9icnZ21pJcCq3mddFRoonPdNpUdSjTSQYh9nJXS0,14977
 indexify/executor/function_executor/task_input.py,sha256=wSrHR4m0juiGClQyeVdhRC37QzDt6Rrjq-ZXJkfBi9k,584
 indexify/executor/function_executor/task_output.py,sha256=vi0W75T8qbxjtAtI-812HksCv986lH_ijEqc3q0CVww,3424
 indexify/executor/grpc/channel_manager.py,sha256=ihDkLoiGBLfSmoA2szbntjCfL3E_NDf5LABRXE7YRec,6330
-indexify/executor/grpc/function_executor_controller.py,sha256=UWjchfRVpmT1uGIJraP25lGDhrKZh1CBtVrnrRK2dXE,16265
+indexify/executor/grpc/function_executor_controller.py,sha256=3esPU9ILzYQ4gjfkRZ3IKHpqXWUQEmNit8XzQYKXoGs,17125
 indexify/executor/grpc/metrics/channel_manager.py,sha256=k-WArgklmP5WhjcmFmrgRblB7yc3XlaOXO8owRyV-mw,649
 indexify/executor/grpc/metrics/state_reconciler.py,sha256=0aI2IM4XztKxFa7NCxYSLafw_iiej3p07yEiKyewXIM,585
 indexify/executor/grpc/metrics/state_reporter.py,sha256=GggBEjMzQUYIG95LtTS4fUg1u9jYowkaXoUXppAXucs,543
 indexify/executor/grpc/metrics/task_controller.py,sha256=9Nm86nGxL2rZ3rAORB0_CBdO--Fe4MBrewVW4CqGyOU,222
 indexify/executor/grpc/state_reconciler.py,sha256=VYfKiy43XOA4OxbQUGVffIeqUaQQ37P4_aD9b_80Ls0,19264
-indexify/executor/grpc/state_reporter.py,sha256=uIouQMbk47pi90ceixKs7QEo_rjltQc0AfajryqbGzM,10494
+indexify/executor/grpc/state_reporter.py,sha256=xXLX1se8iTVbccXlKNGNupQXER1M3Yd0VfpqIQJQJZg,11361
 indexify/executor/grpc/task_controller.py,sha256=JrRUkVa8pGoYWBg_RVfn1ThvDhYCgJSMACFPUww7Lys,20851
+indexify/executor/host_resources/host_resources.py,sha256=V8g6R4ovgV-4mWKRO9qsS_SRCJScs4bcFdvxhk2Az_8,1560
+indexify/executor/host_resources/nvidia_gpu.py,sha256=EZotTSMJtIQl51CpcbYoZpJcrQq1B6f8AbnxE2VI9to,2595
+indexify/executor/host_resources/nvidia_gpu_allocator.py,sha256=oULSjL0AVo_nqR_pquq17079UalHQkhMwMqf72gbPHo,1872
 indexify/executor/metrics/downloader.py,sha256=lctPh8xjkXeLEFJnl1hNrD1yEhLhIl5sggsR4Yoe_Zc,2746
 indexify/executor/metrics/executor.py,sha256=ua-Vv_k1CB4juJdF7tEBQbBMksqWAA3iXKKMKXZUCLk,2369
 indexify/executor/metrics/task_fetcher.py,sha256=iJEwCLzYr2cuz7hRvNiqaa2nvQP4OrA0hm0iJY0YKG0,736
@@ -55,11 +58,11 @@ indexify/executor/runtime_probes.py,sha256=bo6Dq6AGZpJH099j0DHtVSDEH80tv3j9MXf3V
 indexify/executor/task_fetcher.py,sha256=p3iEsWyGi0ZMPAv0183smzOUD1KycQ_dXsyd9mpB9IU,3529
 indexify/executor/task_reporter.py,sha256=7X-IdLdwNBIfFbazG_4rtfR1A0ZFt03JGYpVJQUTKpE,16704
 indexify/executor/task_runner.py,sha256=UupZbGxU9BN4i1t6M8tH-5k3s4eUPEhMhar1YI0Aztk,7219
-indexify/proto/executor_api.proto,sha256=K1lwFmk042GA1tp8s633FZJVg6Fi8f8LtAuFj8Gz7XU,9930
-indexify/proto/executor_api_pb2.py,sha256=5y570_FIgc6WFhHVAKWFieMuUhyKBA7rPJJ4DJ5hcCM,14054
-indexify/proto/executor_api_pb2.pyi,sha256=5eJJJjPNdTMSttNUOtzGwADbASsCh7138de_Y3l8uq4,18612
+indexify/proto/executor_api.proto,sha256=0l0kDcSNwAU-uTW7GTlg35DjYFs1iCiCVuaGZR1KOoY,10505
+indexify/proto/executor_api_pb2.py,sha256=O-yQZt0jMTVXTTXCDs8qZeIi_fAF7Xi-youCXjjzNsA,15037
+indexify/proto/executor_api_pb2.pyi,sha256=carOjGvH65dc7M50g_PTe6PXIMVpRLM_qWRaY28way4,19629
 indexify/proto/executor_api_pb2_grpc.py,sha256=GGiDtyQlA2382E_ZyKUBYcWNEJHH_RlulieStKfkJXI,9514
-indexify-0.3.22.dist-info/METADATA,sha256=cLQOCV8XooTEuzQDpuDKS-sJ6j78nM_7L889e_YUPCw,1198
-indexify-0.3.22.dist-info/WHEEL,sha256=RaoafKOydTQ7I_I3JTrPCg6kUmTgtm4BornzOqyEfJ8,88
-indexify-0.3.22.dist-info/entry_points.txt,sha256=GU9wmsgvN7nQw3N2X0PMYn1RSvF6CrhH9RuC2D8d3Gk,53
-indexify-0.3.22.dist-info/RECORD,,
+indexify-0.3.24.dist-info/METADATA,sha256=qUoUwLmKXouQwXbukZCYo0naWCySii5uI9GPuNEgeSI,1198
+indexify-0.3.24.dist-info/WHEEL,sha256=RaoafKOydTQ7I_I3JTrPCg6kUmTgtm4BornzOqyEfJ8,88
+indexify-0.3.24.dist-info/entry_points.txt,sha256=GU9wmsgvN7nQw3N2X0PMYn1RSvF6CrhH9RuC2D8d3Gk,53
+indexify-0.3.24.dist-info/RECORD,,