Add memory / CPU usage properties to pipeline nodes

.github/ISSUE_TEMPLATE/pipeline-issue-report.md

@@ -0,0 +1,45 @@
+---
+name: Pipeline issue report
+about: Report a potential issue related to the pipeline editor or pipeline execution
+title: ''
+labels: component:pipeline-editor, component:pipeline-runtime, status:Needs Triage, kind:user-feedback
+assignees: ''
+
+---
+
+**Describe the issue**
+A clear and concise description of what the issue is.
+
+**To Reproduce**
+Steps to reproduce the behavior:
+1. Go to '...'
+2. Click on '....'
+3. Scroll down to '....'
+4. See error
+
+**Screenshots or log output**
+If applicable, add screenshots or log output to help explain your problem.
+<details><summary>Log Output</summary>
+<pre>
+Paste the log output here.
+</pre>
+</details>
+
+**Expected behavior**
+A clear and concise description of what you expected to happen.
+
+**Deployment information**
+Describe what you've deployed and how:
+ - Elyra version: [e.g. 1.5.3]
+ - Operating system: [e.g. macos, linux]
+ - Installation source: [e.g. PyPI, conda, from source, official container image, custom container image]
+ - Deployment type: [e.g. local installation, Docker, Kubernetes, Kubeflow [notebook server] , Open Data Hub]
+
+**Pipeline runtime environment**
+If the issue is related to pipeline execution, identify the environment where the pipeline is executed
+ - Local execution
+ - Kubeflow Pipelines (provide version number, whether multi-user auth enabled)
+ - Apache Airflow (provide version number)
+
+**Runtime configuration settings**
+If the issue is related to pipeline execution, document the runtime configuration settings from the Elyra UI, omitting confidential information. 

elyra/pipeline/parser.py

@@ -133,6 +133,9 @@ def _create_pipeline_operation(node: Dict, super_node: Optional[Dict] = None):
             id=node_id,
             type=node.get('type'),
             classifier=node.get('op'),
+            cpu=PipelineParser._get_app_data_field(node, 'cpu'),
+            gpu=PipelineParser._get_app_data_field(node, 'gpu'),
+            memory=PipelineParser._get_app_data_field(node, 'memory'),
             filename=PipelineParser._get_app_data_field(node, 'filename'),
             runtime_image=PipelineParser._get_app_data_field(node, 'runtime_image'),
             dependencies=PipelineParser._get_app_data_field(node, 'dependencies', []),

elyra/pipeline/pipeline.py

@@ -22,8 +22,8 @@ class Operation(object):
     Represents a single operation in a pipeline
     """
 
-    def __init__(self, id, type, classifier, filename, runtime_image, dependencies=None,
-                 include_subdirectories: bool = False, env_vars=None, inputs=None, outputs=None,
+    def __init__(self, id, type, classifier, filename, runtime_image, memory=None, cpu=None, gpu=None,
+                 dependencies=None, include_subdirectories: bool = False, env_vars=None, inputs=None, outputs=None,
                  parent_operations=None):
         """
         :param id: Generated UUID, 128 bit number used as a unique identifier
@@ -42,6 +42,9 @@ def __init__(self, id, type, classifier, filename, runtime_image, dependencies=N
         :param inputs: List of files to be consumed by this operation, produced by parent operation(s)
         :param outputs: List of files produced by this operation to be included in a child operation(s)
         :param parent_operations: List of parent operation 'ids' required to execute prior to this operation
+        :param cpu: number of cpus requested to run the operation
+        :param memory: amount of memory requested to run the operation (in Gi)
+        :param gpu: number of gpus requested to run the operation
         """
 
         # validate that the operation has all required properties
@@ -67,6 +70,9 @@ def __init__(self, id, type, classifier, filename, runtime_image, dependencies=N
         self._inputs = inputs or []
         self._outputs = outputs or []
         self._parent_operations = parent_operations or []
+        self._cpu = cpu
+        self._gpu = gpu
+        self._memory = memory
 
     @property
     def id(self):
@@ -104,6 +110,18 @@ def include_subdirectories(self):
     def env_vars(self):
         return self._env_vars
 
+    @property
+    def cpu(self):
+        return self._cpu
+
+    @property
+    def memory(self):
+        return self._memory
+
+    @property
+    def gpu(self):
+        return self._gpu
+
     def env_vars_as_dict(self, logger: Optional[object] = None) -> Dict:
         """Operation stores environment variables in a list of name=value pairs, while
            subprocess.run() requires a dictionary - so we must convert.  If no envs are
@@ -155,7 +173,10 @@ def __eq__(self, other: object) -> bool:
                 self.include_subdirectories == other.include_subdirectories and \
                 self.outputs == other.outputs and \
                 self.inputs == other.inputs and \
-                self.parent_operations == other.parent_operations
+                self.parent_operations == other.parent_operations and \
+                self.cpu == other.cpu and \
+                self.gpu == other.gpu and \
+                self.memory == other.memory
 
     def __str__(self) -> str:
         return "componentID : {id} \n " \
@@ -166,15 +187,21 @@ def __str__(self) -> str:
                "filename : {filename} \n " \
                "inputs : {inputs} \n " \
                "outputs : {outputs} \n " \
-               "runtime image : {image} \n ".format(id=self.id,
-                                                    name=self.name,
-                                                    parent_op=self.parent_operations,
-                                                    depends=self.dependencies,
-                                                    inc_subdirs=self.include_subdirectories,
-                                                    filename=self.filename,
-                                                    inputs=self.inputs,
-                                                    outputs=self.outputs,
-                                                    image=self.runtime_image)
+               "image : {image} \n " \
+               "gpu: {gpu} \n " \
+               "memory: {memory} \n " \
+               "cpu : {cpu} \n ".format(id=self.id,
+                                        name=self.name,
+                                        parent_op=self.parent_operations,
+                                        depends=self.dependencies,
+                                        inc_subdirs=self.include_subdirectories,
+                                        filename=self.filename,
+                                        inputs=self.inputs,
+                                        outputs=self.outputs,
+                                        image=self.runtime_image,
+                                        gpu=self.gpu,
+                                        cpu=self.cpu,
+                                        memory=self.memory)
 
 
 class Pipeline(object):

elyra/pipeline/processor_kfp.py

@@ -270,6 +270,9 @@ def _cc_pipeline(self, pipeline, pipeline_name):
                                                     pipeline_outputs=operation.outputs,
                                                     pipeline_envs=pipeline_envs,
                                                     emptydir_volume_size=emptydir_volume_size,
+                                                    cpu_request=operation.cpu,
+                                                    mem_request=operation.memory,
+                                                    gpu_limit=operation.gpu,
                                                     image=operation.runtime_image,
                                                     file_outputs={
                                                         'mlpipeline-metrics':

elyra/templates/kfp_template.jinja2

@@ -22,6 +22,9 @@ def create_pipeline():
                                                       cos_dependencies_archive='{{ operation.cos_dependencies_archive }}',
                                                       pipeline_inputs={{ operation.pipeline_inputs }},
                                                       pipeline_outputs={{ operation.pipeline_outputs }},
+                                                      cpu_request='{{ operation.cpu_request }}',
+                                                      mem_request='{{ operation.mem_request }}G',
+                                                      gpu_limit='{{ operation.gpu_limit }}',
                                                       image='{{ operation.image }}',
                                                       file_outputs={
                                                         'mlpipeline-metrics': '{{ metrics_file }}',

packages/pipeline-editor/src/PipelineEditorWidget.tsx

@@ -492,6 +492,9 @@ export class PipelineEditor extends React.Component<
       app_data.dependencies;
     node_props.parameterDef.current_parameters.include_subdirectories =
       app_data.include_subdirectories;
+    node_props.parameterDef.current_parameters.cpu = app_data.cpu;
+    node_props.parameterDef.current_parameters.memory = app_data.memory;
+    node_props.parameterDef.current_parameters.gpu = app_data.gpu;
     node_props.parameterDef.titleDefinition = {
       title: this.canvasController.getNode(source.id).label,
       editable: true
@@ -540,6 +543,9 @@ export class PipelineEditor extends React.Component<
     app_data.env_vars = propertySet.env_vars;
     app_data.dependencies = propertySet.dependencies;
     app_data.include_subdirectories = propertySet.include_subdirectories;
+    app_data.cpu = propertySet.cpu;
+    app_data.memory = propertySet.memory;
+    app_data.gpu = propertySet.gpu;
     this.validateAllNodes();
     this.updateModel();
   }
@@ -848,6 +854,22 @@ export class PipelineEditor extends React.Component<
     }
   }
 
+  cleanNullProperties(): void {
+    // Delete optional fields that have null value
+    for (const node of this.canvasController.getPipelineFlow().pipelines[0]
+      .nodes) {
+      if (node.app_data.cpu === null) {
+        delete node.app_data.cpu;
+      }
+      if (node.app_data.memory === null) {
+        delete node.app_data.memory;
+      }
+      if (node.app_data.gpu === null) {
+        delete node.app_data.gpu;
+      }
+    }
+  }
+
   async handleExportPipeline(): Promise<void> {
     // Warn user if the pipeline has invalid nodes
     const errorMessage = await this.validatePipeline();
@@ -906,6 +928,8 @@ export class PipelineEditor extends React.Component<
       this.widgetContext.path
     );
 
+    this.cleanNullProperties();
+
     pipelineFlow.pipelines[0]['app_data']['name'] = pipeline_name;
     pipelineFlow.pipelines[0]['app_data']['runtime'] = runtime;
     pipelineFlow.pipelines[0]['app_data']['runtime-config'] = runtime_config;
@@ -1259,6 +1283,8 @@ export class PipelineEditor extends React.Component<
       this.widgetContext.path
     );
 
+    this.cleanNullProperties();
+
     pipelineFlow.pipelines[0]['app_data']['name'] =
       dialogResult.value.pipeline_name;
     pipelineFlow.pipelines[0]['app_data']['runtime'] = runtime;