Checklist

User Stories Documented
User Stories Reviewed
Design Reviewed
APIs reviewed
Release priorities assigned
Test cases reviewed
Blog post

Introduction

CDAP pipeline is composed of various plugins that can be configured by users as CDAP pipelines are being developed. While building CDAP pipelines, pipeline developer can provide invalid plugin configurations or schema. For example, the BigQuery sink plugin can have output schema which does not match with underlying BigQuery table. CDAP pipeline developer can use new validation endpoint to validate the stages before deploying the pipeline. In order to fail fast and for better user experience, validation endpoint should return all the validation errors from a given stage when this endpoint is called.

Data pipeline app exposes various error types for plugin validation. In future releases, new error types can be introduced. With current implementation, when plugins with new error types are pushed to hub, data pipeline artifacts need to be updated for every new type of error that is introduced. This is because the validation errors are defined in the data pipeline app itself. A better approach would be to modify data pipeline app so that app artifacts do not need to be replaced for every new type of error.

Goals

To fail fast and for better user experience, introduce a new api to collect multiple error messages from a stage at configure time
Decouple validation error types from data pipeline app
Instrument plugins to use this api to return multiple error messages for validation endpoint

User Stories

As a CDAP pipeline developer, when I validate a stage, I expect that all the invalid config properties and input/output schema fields are highlighted on CDAP UI with appropriate error message.
As a plugin developer, I should be able to capture all the validation errors while configuring the plugin so that all the validation errors can be surfaced on CDAP UI.
As a plugin developer, I should be able to use new validation error types without replacing data pipeline app artifacts.

API Changes for Plugin Validation

Collect Multiple errors from plugins

To collect multiple stage validation errors from the stage, StageConfigurer, MultiInputStageConfigurer and MultiOutputStageConfigurer can be modified as below. If there are any validation errors added to stage configurer, the pipeline deployment will fail and all the errors will be returned as a response to stage validation REST endpoint.

Current implementation does not expose stage name to the plugin in configurePipeline method. Stage name will be needed by the plugins to create stage specific errors. For that stageName will be exposed to through stage configurer as below.

StageConfigurer.java

public interface StageConfigurer {

  ...

/**
 * get the stage name.
 * @return stage name
 */
String getStageName();

/**
 * add validation errors.
 * @param e errors
 */
void addValidationError(ValidationError e);

Decouple plugin error types from data pipeline app

Approach - 1

A new ValidationError class is introduced to collect multiple validation errors in stage configurer. This class will be exposed by data pipeline app to the plugins. Each new type of error can be added to hydrator-common which can be added as compile time dependency to the plugins. This approach allows us to easily add more types of validation errors for plugins while removing a need to update data pipeline artifact for each new error type.

ValidationError.java

/**
 * Represents error that occurred during validation.
 */
@Beta
public class ValidationError {
  private final String message;
  private final String type;

  /**
   * Creates an error with provided error message.
   * @param message error message
   */
  public ValidationError(String message) {
    this.message = message;
    this.type = getType();
  }

  /**
   * Returns the type of the error.
   */
  public String getType() {
    return "ERROR";
  }

  @Override
  public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    ValidationError error = (ValidationError) o;
    return message.equals(error.message) &&
      type.equals(error.type);
  }

  @Override
  public int hashCode() {
    return Objects.hash(message, type);
  }
}

Introduced Errors

Below are some of the error types that will be added to hydrator-common module. Note that this list will keep evolving as new types of errors are added to the module.

InvalidStageError.java

/**
 * Represents error that occurred during stage validation.
 */
@Beta
public class InvalidStageError extends ValidationError {
  private final String stage;

  /**
   * Creates error that represents a stage validation error.
   *
   * @param message error message
   * @param stage name of the stage
   */
  public InvalidStageError(String message, String stage) {
    super(message);
    this.stage = stage;
  }

  @Override
  public String getType() {
    return "STAGE_ERROR";
  }

  @Override
  public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    if (!super.equals(o)) {
      return false;
    }
    InvalidStageError that = (InvalidStageError) o;
    return Objects.equals(stage, that.stage);
  }

  @Override
  public int hashCode() {
    return Objects.hash(super.hashCode(), stage);
  }
}

InvalidStagePropertyError.java

/**
 * Represents invalid stage property error that occurred during stage validation.
 */
@Beta
public class InvalidStagePropertyError extends InvalidStageError {
  private final String property;

  /**
   * Creates error that represents invalid stage property.
   *
   * @param message error message
   * @param stage name of the stage
   * @param property invalid property
   */
  public InvalidStagePropertyError(String message, String stage, String property) {
    super(message, stage);
    this.property = property;
  }

  @Override
  public String getType() {
    return "INVALID_FIELD";
  }

  @Override
  public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    if (!super.equals(o)) {
      return false;
    }
    InvalidStagePropertyError that = (InvalidStagePropertyError) o;
    return property.equals(that.property);
  }

  @Override
  public int hashCode() {
    return Objects.hash(super.hashCode(), property);
  }
}

InvalidInputSchemaFieldError.java

/**
 * Represents invalid schema field in input schema.
 */
public class InvalidInputSchemaFieldError extends InvalidStageError {
  private final String field;

  /**
   * Creates error that represents invalid schema field in input schema.
   * @param message error message
   * @param stage name of the stage
   * @param field field that is invalid in input schema
   */
  public InvalidInputSchemaFieldError(String message, String stage, String field) {
    super(message, stage);
    this.field = field;
  }

  @Override
  public String getType() {
    return "INVALID_INPUT_SCHEMA";
  }

  @Override
  public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    if (!super.equals(o)) {
      return false;
    }
    InvalidInputSchemaFieldError that = (InvalidInputSchemaFieldError) o;
    return field.equals(that.field);
  }

  @Override
  public int hashCode() {
    return Objects.hash(super.hashCode(), field);
  }
}

InvalidOutputSchemaFieldError.java

/**
 * Represents invalid schema field in output schema.
 */
public class InvalidOutputSchemaFieldError extends InvalidStageError {
  private final String field;

  /**
   * Creates error that represents invalid schema field in output schema.
   * @param message error message
   * @param stage name of the stage
   * @param field field that is invalid in output schema
   */
  public InvalidOutputSchemaFieldError(String message, String stage, String field) {
    super(message, stage);
    this.field = field;
  }

  @Override
  public String getType() {
    return "INVALID_OUTPUT_SCHEMA";
  }

  @Override
  public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    if (!super.equals(o)) {
      return false;
    }
    InvalidOutputSchemaFieldError that = (InvalidOutputSchemaFieldError) o;
    return field.equals(that.field);
  }

  @Override
  public int hashCode() {
    return Objects.hash(super.hashCode(), field);
  }
}

API usage in plugins

@Override
public void configurePipeline(PipelineConfigurer pipelineConfigurer) {
  pipelineConfigurer.createDataset(conf.destinationFileset, FileSet.class);
  StageConfigurer stageConfigurer = pipelineConfigurer.getStageConfigurer();
  // get the name of the stage 
  String stageName = stageConfigurer.getStageName();
  try {
    Pattern.compile(conf.filterRegex);
  } catch (Exception e) {  
    // add validation error to stage configurer
    stageConfigurer.addValidationError(new InvalidStagePropertyError(e.getMessage(), stageName, "filterRegex"));
  }
  if (conf.sourceFileset.equals(conf.destinationFileset)) {
    // add validation error to stage configurer
    stageConfigurer.addValidationError(new InvalidStageError("source and destination filesets must be different", stageName));
  }
}

Approach - 2

Validation error represents an error with various causes with extra properties that caused the error. For example, when the input schema field type does not match the underlying sink schema, information such as error message, stage name, field name, suggested type will be useful information to have in the validation error. Each error message can be associated to more than one causes. This can happen for plugins such as joiner and splitter where there are multiple input or output schemas from a given stage. For example, when input schemas for joiner are not compatible, the validation error information should contain error message along with the fields from both the stages that caused this error. This means that a validation error can be represented as a list of causes where each cause is a map of error cause to its value as shown below.

ValidationError.java

/**
 * Represents error that occurred during validation.
 */
@Beta
public class ValidationError {
  private final String message;
  protected final List<Map<String, Object>> causes;

  /**
   * Creates a validation error with a message and empty map of causes
   * @param message
   */
  public ValidationError(String message) {
    this.message = message;	
    this.causes = new ArrayList<>();
  }
 
  @Override
   public boolean equals(Object o) {
    if (this == o) {
      return true;
    }
    if (o == null || getClass() != o.getClass()) {
      return false;
    }
    ValidationError that = (ValidationError) o;
    return message.equals(that.message) && causes.equals(that.causes);
  }

  @Override
  public int hashCode() {
    return Objects.hash(message, causes);
  }
  
}

All the attributes of an error can be tracked at central location as below:

ErrorAttributes.java

/**
 * Error attributes.
 */
public enum ErrorAttributes {
  STAGE("stage"), // represents stage being validated
  PROPERTY("property"), // represents stage property
  INPUT_FIELD("input_field") // represents field in the input schema
  OUTPUT_FIELD("output_field") // represents field in the output schema
  OUTPUT_PORT("output_port"), // represents output port for plugins such as SplitterTransform where multiple output schemas are expected
  INPUT_STAGE("input_stage"), // represents input stage for plugins such as Joiner where multiple input schemas are expected
  ..

  private String name;

  ErrorAttributes(String name) {
    this.name = name;
  }
}

Introduced Errors

With this approach following error classes can be added to hydrator-common which represents specific type of errors.

InvalidStageError.java

/**
 * Represents error that occurred during stage validation.
 */
@Beta
public class InvalidStageError extends ValidationError {
  /**
   * Creates validation error that occurred during stage validation.
   * @param message error message
   * @param stage name of the stage that caused this validation error
   */
  public InvalidStageError(String message, String stage) {
    super(message);
    causes.add(Collections.singletonMap("stage", stage));
  }
}

InvalidStagePropertyError.java

/**
 * Represents error that occurred during stage config property validation.
 */
@Beta
public class InvalidStagePropertyError extends ValidationError {
 /**
   * Creates validation error that occurred during stage validation.
   * @param message error message
   * @param stage name of the stage that caused this validation error
   * @param property property that is invalid
   */
  public InvalidStageError(String message, String stage, String property) {
    super(message);
    Map<String, Object> map = new HashMap<>();
    map.put("stage", stage);
    map.put("property", property);
    causes.add(map);
  }
}

/**
 * Represents invalid input schema error. 
 */
public class InvalidInputSchemaError extends NewValidationError {

  /**
   * Creates invalid input schema error.
   * @param message error message
   * @param stage name of the stage 
   * @param map map of incoming stage name to field that is invalid.
   */
  public InvalidInputSchemaError(String message, String stage, Map<String, String> map) {
    super(message);
    for (Map.Entry<String, String> entry : map.entrySet()) {
      Map<String, Object> causeMap = new HashMap<>();
      causeMap.put("stage", stage);
      causeMap.put("input_stage", entry.getKey());
      causeMap.put("input_field", entry.getValue());
      causes.add(causeMap);
    }
  }
}

InvalidOutputSchemaError.java

/**
 * Represents invalid output schema error.
 */
public class InvalidOutputSchemaError extends NewValidationError {

  /**
   * Creates invalid output schema error.
   * @param message error message
   * @param stage name of the stage
   * @param map map of output going port name to field that is invalid
   */
  public InvalidOutputSchemaError(String message, String stage, Map<String, String> map) {
    super(message);
    for (Map.Entry<String, String> entry : map.entrySet()) {
      Map<String, Object> causeMap = new HashMap<>();
      causeMap.put("stage", stage);
      causeMap.put("output_port", entry.getKey());
      causeMap.put("output_field", entry.getValue());
      causes.add(causeMap);
    }
  }
}

API usage in plugins

The above api usage in the plugins is same as approach-1.

@Override
public void configurePipeline(PipelineConfigurer pipelineConfigurer) {
  pipelineConfigurer.createDataset(conf.destinationFileset, FileSet.class);
  StageConfigurer stageConfigurer = pipelineConfigurer.getStageConfigurer();
  // get the name of the stage 
  String stageName = stageConfigurer.getStageName();
  try {
    Pattern.compile(conf.filterRegex);
  } catch (Exception e) {  
    // add validation error to stage configurer
    stageConfigurer.addValidationError(new InvalidStagePropertyError(e.getMessage(), stageName, "filterRegex"));
  }
  if (conf.sourceFileset.equals(conf.destinationFileset)) {
    // add validation error to stage configurer
    stageConfigurer.addValidationError(new InvalidStageError("source and destination filesets must be different", stageName));
  }
}

Impact on UI

Type	Description	Scenario	Approach - 1 - Json Response	Approach - 2 - Json Response
STAGE_ERROR	Represents validation error while configuring the stage	If there is any error while connecting to sink while getting actual schema	{ "errors": [ { "type" : "STAGE_ERROR", "stage" : "src", `"message"` `: "Could not load jdbc driver."` } ] }	{ "errors": [ { "message": "Could not load jdbc driver.", "causes": [ { "stage": "src" } ] } ] }
INVALID_PROPERTY	Represents invalid configuration property	If config property value contains characters that are not allowed by underlying source or sink	{ "errors": [ { "type" : "INVALID_PROPERTY", "stage" : "projection", "message" : "Can not specify both drop and keep. One should be empty or null", "property" : "drop" }, { "type" : "INVALID_PROPERTY", "stage" : "projection", "message" : "Can not specify both drop and keep. One should be empty or null", "property" : "keep" } ] }	{ "errors": [ { "message": "Can not specify both drop and keep. One should be empty or null", "causes": [ { "stage": "projection", "property": "keep" }, { "stage" : "projection", "property" : "drop" } ] } ] }
PLUGIN_NOT_FOUND	Represents plugin not found error for a stage	If the plugin was not found. This error will be thrown from the data pipeline app	{ "errors": [ { "stage": "src", "type": "PLUGIN_NOT_FOUND", "message": "Plugin named 'Mock' of type 'batchsource' not found.", "pluginType": "batchsource", "pluginName": "Mock", "requestedArtifact": { "scope": "USER", "name": "app-mocks-ghost", "version": "1.0.0" } } ] }	{ "errors": [ { "message": "Plugin named 'Mock' of type 'batchsource' not found.", "causes": [ { "stage": "src", "pluginType": "batchsource", "pluginName": "Mock", "requestedArtifact": { "scope": "USER", "name": "app-mocks-ghost", "version": "1.0.0" } } ] } ] }
INVALID_INPUT_SCHEMA	Represents invalid schema field in input schema	If the input schemas for joiner plugin is of different types	{ "errors": [ { "type" : "INVALID_INPUT_SCHEMA", "stage" : "joiner", "message" : "Invalid schema field 'id'. Different types of join keys found.", "field" : "id" } ] }	{ "errors": [ { "message": "Invalid schema field 'id'. Different types of join keys found.", "causes": [ { "stage": "joiner", "input_stage": "source1", "input_field": "id" }, { "stage": "joiner", "input_stage": "source2", "input_field": "id" } ] } ] }
INVALID_OUTPUT_SCHEMA	Represents invalid schema field in output schema	If the output schema for the plugin is not compatible with underlying sink	{ "errors": [ { "type" : "INVALID_OUTPUT_SCHEMA", "stage" : "splitter", "message" : "Invalid schema field 'email'. It should be of type 'string' at output port 'port'", "field" : "email" } ] }	{ "errors": [ { "message": "Invalid schema field 'email'. It should be of type 'string'", "causes": [ { "stage": "splitter", "output_port": "port", "output_field": "email" } ] } ] }

Conclusion

As it is observed in the above json responses, approach 1 will require specific classes for each error attributes. However, that approach 2 is more flexible than approach 1 in terms of what error classes can have as attributes. Since approach 2 avoids class explosion, that is the suggested approach.

Related Jira

Error rendering macro 'jira' : Unable to locate Jira server for this macro. It may be due to Application Link configuration.