Error Codes

Complete reference of 33 error codes across 11 categories returned by the EventML Service.

Error response formatjson

{
  "error": {
    "code": "ML_1000",
    "message": "Model configuration is invalid or missing required fields",
    "http": 400
  }
}

ModelsML_1000-1099

Code	HTTP	Name	Description	Resolution
`ML_1000`	400	`INVALID_MODEL_CONFIG`	Model configuration is invalid or missing required fields	Check model framework, version, and required fields against API docs
`ML_1001`	404	`MODEL_NOT_FOUND`	Specified model does not exist	Verify model ID via GET /models
`ML_1002`	409	`MODEL_ALREADY_EXISTS`	A model with this name already exists for this tenant	Use a different name or update the existing model
`ML_1003`	400	`INVALID_FRAMEWORK`	Unsupported ML framework specified	Use: PYTORCH, TENSORFLOW, SKLEARN, XGBOOST, ONNX, or CUSTOM

Code	HTTP	Name	Description	Resolution
`ML_1100`	400	`INVALID_TRAINING_CONFIG`	Training job configuration is invalid	Check hyperparameters, data config, and instance settings
`ML_1101`	404	`TRAINING_JOB_NOT_FOUND`	Training job does not exist	Verify job ID via GET /training
`ML_1102`	409	`TRAINING_ALREADY_RUNNING`	A training job is already running for this model version	Stop the existing job before starting a new one
`ML_1103`	502	`TRAINING_INFRASTRUCTURE_ERROR`	Training infrastructure (SageMaker/EC2) returned an error	Retry; infrastructure may be scaling up
`ML_1104`	504	`TRAINING_TIMEOUT`	Training job exceeded maximum allowed duration	Increase maxRuntime or reduce dataset size

Code	HTTP	Name	Description	Resolution
`ML_1200`	400	`INVALID_INFERENCE_INPUT`	Inference input format does not match model schema	Check input schema for the model version endpoint
`ML_1201`	404	`ENDPOINT_NOT_FOUND`	Inference endpoint does not exist	Create an endpoint via POST /endpoints
`ML_1202`	503	`ENDPOINT_NOT_READY`	Endpoint is not in ACTIVE state (still provisioning or scaling)	Wait for endpoint to reach ACTIVE status before invoking
`ML_1203`	504	`INFERENCE_TIMEOUT`	Inference request timed out before returning a prediction	Reduce input size or increase endpoint timeout configuration

Code	HTTP	Name	Description	Resolution
`ML_1300`	400	`INVALID_EXPERIMENT_CONFIG`	Experiment configuration is invalid	Ensure variant traffic allocations sum to 100%
`ML_1301`	404	`EXPERIMENT_NOT_FOUND`	Experiment does not exist	Verify experiment ID via GET /experiments
`ML_1302`	409	`EXPERIMENT_ALREADY_RUNNING`	Experiment is already in RUNNING state	Stop the experiment before modifying its configuration

Code	HTTP	Name	Description	Resolution
`ML_1400`	400	`INVALID_DRIFT_CONFIG`	Drift monitor configuration is invalid	Check threshold, detection method (KS_TEST, PSI, etc.), and feature list
`ML_1401`	404	`DRIFT_MONITOR_NOT_FOUND`	Drift monitor does not exist	Create a monitor via POST /drift-monitors
`ML_1402`	409	`DRIFT_DETECTED`	Significant data drift detected above threshold	Retrain the model with updated data distribution

Code	HTTP	Name	Description	Resolution
`ML_1500`	400	`INVALID_FEATURE_GROUP`	Feature group configuration is invalid	Check feature schema, data types, and entity key configuration
`ML_1501`	404	`FEATURE_GROUP_NOT_FOUND`	Feature group does not exist	Create via POST /feature-store/groups
`ML_1502`	409	`FEATURE_VERSION_CONFLICT`	Feature version already exists with conflicting schema	Use a new version number or update the existing version

Code	HTTP	Name	Description	Resolution
`ML_1600`	403	`GOVERNANCE_VIOLATION`	Model governance policy violated (missing documentation, tests, or approvals)	Review and meet all governance requirements before releasing
`ML_1601`	400	`MISSING_APPROVAL`	Required approval not obtained for this release stage	Submit model for approval workflow via POST /governance/approvals

Code	HTTP	Name	Description	Resolution
`ML_1700`	401	`UNAUTHORIZED`	Authentication required	Provide a valid JWT token in the Authorization header
`ML_1701`	403	`FORBIDDEN`	Insufficient permissions for this ML operation	Request required role (ml-engineer, admin) from your admin
`ML_1702`	401	`INVALID_TOKEN`	Invalid or expired authentication token	Obtain a fresh JWT token via auth-svc /login or /refresh

Code	HTTP	Name	Description	Resolution
`ML_1800`	400	`VALIDATION_FAILED`	Request body validation failed	Check request body against the API docs schema
`ML_1801`	400	`INVALID_INPUT`	Invalid input provided for a field	Review field constraints (type, min/max, pattern, enum values)

Code	HTTP	Name	Description	Resolution
`ML_1900`	500	`DATABASE_ERROR`	Database operation failed	Retry; Aurora Serverless v2 may be scaling up from idle
`ML_1901`	500	`TRANSACTION_FAILED`	Database transaction failed due to conflict or timeout	Retry the operation with a new Idempotency-Key

Code	HTTP	Name	Description	Resolution
`ML_2000`	500	`INTERNAL_SERVER_ERROR`	Internal server error	Retry; if persistent, contact support with the x-request-id value
`ML_2001`	503	`SERVICE_UNAVAILABLE`	Service temporarily unavailable (ECS task restarting)	Retry with exponential backoff
`ML_2002`	504	`TIMEOUT`	Request timeout exceeded	Reduce request complexity or retry