Skip to main content
Version: Next

Configuration

We recommend configuring the client with an openlineage.yml file that contains all the details of how to connect to your OpenLineage backend.

See example configurations.

You can make this file available to the client in three ways (the list also presents precedence of the configuration):

  1. Set an OPENLINEAGE_CONFIG environment variable to a file path: OPENLINEAGE_CONFIG=path/to/openlineage.yml.
  2. Place an openlineage.yml in the user's current working directory.
  3. Place an openlineage.yml under .openlineage/ in the user's home directory (~/.openlineage/openlineage.yml).

Environment Variables​

The following environment variables are available:

NameDescriptionSince
OPENLINEAGE_CONFIGThe path to the YAML configuration file. Example: path/to/openlineage.yml
OPENLINEAGE_DISABLEDWhen true, OpenLineage will not emit events.0.9.0

You can also configure the client with dynamic environment variables.

The OpenLineage client supports configuration through dynamic environment variables.

Configuring OpenLineage Client via Dynamic Environment Variables​

These environment variables must begin with OPENLINEAGE__, followed by sections of the configuration separated by a double underscore __. All values in the environment variables are automatically converted to lowercase, and variable names using snake_case (single underscore) are converted into camelCase within the final configuration.

Key Features​

  1. Prefix Requirement: All environment variables must begin with OPENLINEAGE__.
  2. Sections Separation: Configuration sections are separated using double underscores __ to form the hierarchy.
  3. Lowercase Conversion: Environment variable values are automatically converted to lowercase.
  4. CamelCase Conversion: Any environment variable name using single underscore _ will be converted to camelCase in the final configuration.
  5. JSON String Support: You can pass a JSON string at any level of the configuration hierarchy, which will be merged into the final configuration structure.
  6. Hyphen Restriction: You cannot use - in environment variable names. If a name strictly requires a hyphen, use a JSON string as the value of the environment variable.
  7. Precedence Rules:
  • Top-level keys have precedence and will not be overwritten by more nested entries.
  • For example, OPENLINEAGE__TRANSPORT='{..}' will not have its keys overwritten by OPENLINEAGE__TRANSPORT__AUTH__KEY='key'.

Examples​

Setting following environment variables:

OPENLINEAGE__TRANSPORT__TYPE=http
OPENLINEAGE__TRANSPORT__URL=http://localhost:5050
OPENLINEAGE__TRANSPORT__ENDPOINT=/api/v1/lineage
OPENLINEAGE__TRANSPORT__AUTH__TYPE=api_key
OPENLINEAGE__TRANSPORT__AUTH__API_KEY=random_token
OPENLINEAGE__TRANSPORT__COMPRESSION=gzip

is equivalent to passing following YAML configuration:

transport:
  type: http
  url: http://localhost:5050
  endpoint: api/v1/lineage
  auth:
    type: api_key
    apiKey: random_token
  compression: gzip

Facets Configuration​

In YAML configuration file you can also disable facets to filter them out from the OpenLineage event.

YAML Configuration

transport:
  type: console
facets:
  spark_unknown:
    disabled: true
  "spark.logicalPlan":
    disabled: true

Deprecated and removed syntax​

The following syntax was deprecated and got removed:

facets:
  disabled:
    - spark_unknown
    - spark.logicalPlan

Please be aware that this syntax is not working anymore.

Transports​

Tip: See current list of all supported transports.

HTTP​

Allows sending events to HTTP endpoint, using ApacheHTTPClient.

Configuration​

  • type - string, must be "http". Required.
  • url - string, base url for HTTP requests. Required.
  • endpoint - string specifying the endpoint to which events are sent, appended to url. Optional, default: /api/v1/lineage.
  • urlParams - dictionary specifying query parameters send in HTTP requests. Optional.
  • timeoutInMillis - integer specifying timeout (in milliseconds) value used while connecting to server. Optional, default: 5000.
  • auth - dictionary specifying authentication options. Optional, by default no authorization is used. If set, requires the type property.
    • type - string specifying value for one of the out-of-the-box available authentication methods (apiKey or jwt), or the fully qualified class name of your TokenProvider. Required if auth is provided.
    • Configuration options for api_key authentication:
      • apiKey - string setting the Authentication HTTP header as the Bearer. Required if type is api_key.
    • Configuration options for jwt authentication are documented in the JWT Token Provider section.
  • headers - dictionary specifying HTTP request headers. Optional.
  • compression - string, name of algorithm used by HTTP client to compress request body. Optional, default value null, allowed values: gzip. Added in v1.13.0.

Behavior​

Events are serialized to JSON, and then are send as HTTP POST request with Content-Type: application/json.

Examples​

Anonymous connection:

transport:
  type: http
  url: http://localhost:5000

With authorization:

transport:
  type: http
  url: http://localhost:5000
  auth:
    type: api_key
    api_key: f38d2189-c603-4b46-bdea-e573a3b5a7d5

Full example:

transport:
  type: http
  url: http://localhost:5000
  endpoint: /api/v1/lineage
  urlParams:
    param0: value0
    param1: value1
  timeoutInMillis: 5000
  auth:
    type: api_key
    api_key: f38d2189-c603-4b46-bdea-e573a3b5a7d5
  headers:
    X-Some-Extra-Header: abc
  compression: gzip

JWT Token Provider​

The JwtTokenProvider is an authentication provider that exchanges an API key for a JWT token via a POST endpoint. This is useful for services that require OAuth-style authentication where you need to obtain a token before making API requests.

Configuration​

When using JWT authentication with HTTP transport, configure the auth section as follows:

  • type - string, must be "jwt". Required.
  • apiKey - string, the API key used to obtain the JWT token. Required.
  • tokenEndpoint - string, the URL endpoint for token generation. Required.
  • tokenFields - array of strings, JSON field names to search for the token in the response. The provider tries each field in order. Optional, default: ["token", "access_token"].
  • expiresInField - string, JSON field name containing the token expiration time in seconds. Optional, default: "expires_in".
  • grantType - string, OAuth grant type parameter sent in the token request. Optional, default: "urn:ietf:params:oauth:grant-type:jwt-bearer".
  • responseType - string, OAuth response type parameter sent in the token request. Optional, default: "token".
  • tokenRefreshBuffer - integer, number of seconds before token expiry to trigger a refresh. Optional, default: 120.
Behavior​
  • The provider sends a POST request with URL-encoded form data containing the API key and OAuth parameters.
  • The response is expected to be JSON containing the JWT token and optionally an expiration time.
  • Tokens are cached and automatically refreshed before expiration (default: 120 seconds before expiry, configurable via tokenRefreshBuffer).
  • If no expiration is provided in the response, the provider attempts to extract it from the JWT payload's exp claim.
  • The provider supports multiple JSON field names for the token, trying each in order until a match is found.
  • Field matching is case-insensitive and handles both snake_case and camelCase variations (e.g., expires_in matches expiresIn).
Examples​

Standard OAuth configuration:

transport:
  type: http
  url: https://api.example.com
  auth:
    type: jwt
    apiKey: your-api-key
    tokenEndpoint: https://auth.example.com/token

With custom field names:

transport:
  type: http
  url: https://api.example.com
  auth:
    type: jwt
    apiKey: your-api-key
    tokenEndpoint: https://auth.example.com/token
    tokenFields: ["access_token", "token"]
    expiresInField: expires_in

IBM Cloud IAM configuration:

transport:
  type: http
  url: https://api.example.com
  auth:
    type: jwt
    apiKey: your-ibm-api-key
    tokenEndpoint: https://iam.cloud.ibm.com/identity/token
    grantType: urn:ibm:params:oauth:grant-type:apikey
    responseType: cloud_iam

Kafka​

If a transport type is set to kafka, then the below parameters would be read and used when building KafkaProducer. This transport requires the artifact org.apache.kafka:kafka-clients:3.1.0 (or compatible) on your classpath.

Configuration​

  • type - string, must be "kafka". Required.

  • topicName - string specifying the topic on what events will be sent. Required.

  • properties - a dictionary containing a Kafka producer config as in Kafka producer config. Required.

  • localServerId - deprecated, renamed to messageKey since v1.13.0.

  • messageKey - string, key for all Kafka messages produced by transport. Optional, default value described below. Added in v1.13.0.

    Default values for messageKey are:

    • run:{rootJob.namespace}/{rootJob.name} - for RunEvent with parent facet containing link to root job
    • run:{parentJob.namespace}/{parentJob.name} - for RunEvent with parent facet
    • run:{job.namespace}/{job.name} - for RunEvent
    • job:{job.namespace}/{job.name} - for JobEvent
    • dataset:{dataset.namespace}/{dataset.name} - for DatasetEvent

Behavior​

Events are serialized to JSON, and then dispatched to the Kafka topic.

Notes​

It is recommended to provide messageKey if Job hierarchy is used. It can be any string, but it should be the same for all jobs in hierarchy, like Airflow task -> Spark application -> Spark task runs.

Examples​

transport:
  type: kafka
  topicName: openlineage.events
  properties:
    bootstrap.servers: localhost:9092,another.host:9092
    acks: all
    retries: 3
    key.serializer: org.apache.kafka.common.serialization.StringSerializer
    value.serializer: org.apache.kafka.common.serialization.StringSerializer
  messageKey: some-value

Notes: It is recommended to provide messageKey if Job hierarchy is used. It can be any string, but it should be the same for all jobs in hierarchy, like Airflow task -> Spark application.

Default values are:

  • run:{rootJob.namespace}/{rootJob.name} - for RunEvent with parent facet containing link to root job
  • run:{parentJob.namespace}/{parentJob.name}/{parentRun.id} - for RunEvent with parent facet
  • run:{job.namespace}/{job.name}/{run.id} - for RunEvent
  • job:{job.namespace}/{job.name} - for JobEvent
  • dataset:{dataset.namespace}/{dataset.name} - for DatasetEvent

Console​

This straightforward transport emits OpenLineage events directly to the console through a logger. No additional configuration is required.

Behavior​

Events are serialized to JSON. Then each event is logged with INFO level to logger with name ConsoleTransport.

Notes​

Be cautious when using the DEBUG log level, as it might result in double-logging due to the OpenLineageClient also logging.

Configuration​

  • type - string, must be "console". Required.

Examples​

transport:
  type: console

File​

Designed mainly for integration testing, the FileTransport emits OpenLineage events to a given file.

Configuration​

  • type - string, must be "file". Required.
  • location - string specifying the path of the file. Required.

Behavior​

  • If the target file is absent, it's created.
  • Events are serialized to JSON, and then appended to a file, separated by newlines.
  • Intrinsic newline characters within the event JSON are eliminated to ensure one-line events.

Notes for Yarn/Kubernetes​

This transport type is pretty useless on Spark/Flink applications deployed to Yarn or Kubernetes cluster:

  • Each executor will write file to a local filesystem of Yarn container/K8s pod. So resulting file will be removed when such container/pod is destroyed.
  • Kubernetes persistent volumes are not destroyed after pod removal. But all the executors will write to the same network disk in parallel, producing a broken file.

Examples​

transport:
  type: file
  location: /path/to/your/file

Composite​

The CompositeTransport is designed to combine multiple transports, allowing event emission to several destinations. This is useful when events need to be sent to multiple targets, such as a logging system and an API endpoint. The events are delivered sequentially - one after another in a defined order.

Configuration​

  • type - string, must be "composite". Required.
  • transports - a list or a map of transport configurations. Required.
  • continueOnFailure - boolean flag, determines if the process should continue even when one of the transports fails. Default is true.
  • withThreadPool - boolean flag, determines if a thread pool for parallel event emission should be kept between event emissions. Default is true.

Behavior​

  • The configured transports will be initialized and used in sequence (sorted by transport name) to emit OpenLineage events.
  • If continueOnFailure is set to false, a failure in one transport will stop the event emission process, and an exception will be raised.
  • If continueOnFailure is true, the failure will be logged, but the remaining transports will still attempt to send the event.

Notes for Multiple Transports​

The composite transport can be used with any OpenLineage transport (e.g. HttpTransport, KafkaTransport, etc). Ideal for scenarios where OpenLineage events need to reach multiple destinations for redundancy or different types of processing.

The transports configuration can be provided in two formats:

  1. A list of transport configurations, where each transport may optionally include a name field.
  2. A map of transport configurations, where the key acts as the name for each transport. The map format is particularly useful for configurations set via environment variables or Java properties, providing a more convenient and flexible setup.
Why are transport names used?​

Transport names are not required for basic functionality. Their primary purpose is to enable configuration of composite transports via environment variables, which is only supported when names are defined.

Examples​

transport:
  type: composite
  continueOnFailure: true
  transports:
    - type: http
      url: http://example.com/api
      name: my_http
    - type: kafka
      topicName: openlineage.events
      properties:
        bootstrap.servers: localhost:9092,another.host:9092
        acks: all
        retries: 3
        key.serializer: org.apache.kafka.common.serialization.StringSerializer
        value.serializer: org.apache.kafka.common.serialization.StringSerializer
      messageKey: some-value
      continueOnFailure: true

Transform​

The TransformTransport is designed to enable event manipulation before emitting the event. Together with CompositeTransport, it can be used to send different events into multiple backends.

Configuration​

  • type - string, must be "transform". Required.
  • transformerClass - class name of the event transformer. Class has to implement io.openlineage.client.transports.transform.EventTransformer interface and provide public no-arg constructor. Class needs to be available on the classpath. Required.
  • transformerProperties - Extra properties that can be passed into transformerClass based on the configuration. Optional.
  • transport - Transport configuration to emit modified events. Required.

Behavior​

  • The configured transformerClass will be used to alter events before the emission.
  • Modified events will be passed into the configured transport for further processing.
  • In case of returning null, the event will be skipped.

EventTransformer interface​

public class CustomEventTransformer implements EventTransformer {
  @Override
  public void initialize(Map<String, String> properties) { ... }

  @Override
  public RunEvent transform(RunEvent event) { ... }

  @Override
  public DatasetEvent transform(DatasetEvent event) { .. }

  @Override
  public JobEvent transform(JobEvent event) { ... }
}

Examples​

transport:
  type: transform
  transformerClass: io.openlineage.CustomEventTransformer
  transformerProperties:
    key1: value1
    key2: value2
  transport:
    type: http
    url: http://example.com/api
    name: my_http

GcpLineage​

To use this transport in your project, you need to include io.openlineage:transports-gcplineage artifact in your build configuration. This is particularly important for environments like Spark, where this transport must be on the classpath for lineage events to be emitted correctly.

Configuration​

  • type - string, must be "gcplineage". Required.
  • endpoint - string, specifies the endpoint to which events are sent, default value is datalineage.googleapis.com:443. Optional.
  • projectId - string, the project quota identifier. If not provided, it is determined based on user credentials. Optional.
  • location - string, Dataplex location. Optional, default: "us".
  • credentialsFile - string, path to the Service Account credentials JSON file. Optional, if not provided Application Default Credentials are used
  • mode - enum that specifies the type of client used for publishing OpenLineage events to GCP Lineage service. Possible values: sync (synchronous) or async (asynchronous). Optional, default: async.

Behavior​

  • Events are serialized to JSON, included as part of a gRPC request, and then dispatched to the GCP Lineage service endpoint.
  • Depending on the mode chosen, requests are sent using either a synchronous or asynchronous client.

Examples​

transport:
  type: gcplineage
  projectId: your_gcp_project_id
  location: us
  mode: sync
  credentialsFile: path/to/credentials.json

Google Cloud Storage​

To use this transport in your project, you need to include io.openlineage:transports-gcs artifact in your build configuration. This is particularly important for environments like Spark, where this transport must be on the classpath for lineage events to be emitted correctly.

Configuration​

  • type - string, must be "gcs". Required.
  • projectId - string, the project quota identifier. Required.
  • credentialsFile - string, path to the Service Account credentials JSON file. Optional, if not provided Application Default Credentials are used
  • bucketName - string, the GCS bucket name. Required
  • fileNamePrefix - string, prefix for the event file names. Optional.

Behavior​

  • Events are serialized to JSON and stored in the specified GCS bucket.
  • Each event file is named based on its eventTime, converted to epoch milliseconds, with an optional prefix if configured.
  • Two constructors are available: one accepting both Storage and GcsTransportConfig and another solely accepting GcsTransportConfig.

Examples​

transport:
  type: gcs
  bucketName: my-gcs-bucket
  fileNamePrefix: /file/name/prefix/
  credentialsFile: path/to/credentials.json

DataZone Transport​

To use this transport in your project, you need to include io.openlineage:transports-datazone artifact in your build configuration. This is particularly important for environments like Spark, where this transport must be on the classpath for lineage events to be emitted correctly.

Configuration​

  • type - string, must be "amazon_datazone_api". Required.
  • domainId - string, specifies the DataZone / SageMaker Unified Studio domain id. The lineage events will be then sent to the following domain. Required.
  • region - string. When provided, the DataZone client will be configured to use this specific region. If endpointOverride is also provided, this value is not used. Optional, default: None (uses AWS SDK default region resolution).
  • endpointOverride - string, overrides the default HTTP endpoint for Amazon DataZone client. Default value will be set by AWS SDK to following endpoints based on the region. Optional, default: None

Behavior​

  • Events are serialized to JSON, and then dispatched to the DataZone / SageMaker Unified Studio endpoint.

Examples​

transport:
  type: amazon_datazone_api
  domainId: dzd-domain-id

S3​

To use this transport in your project, you need to include the following dependency in your build configuration. This is particularly important for environments like Spark, where this transport must be on the classpath for lineage events to be emitted correctly.

Maven​


<dependency>
    <groupId>io.openlineage</groupId>
    <artifactId>transports-s3</artifactId>
    <version>1.50.0</version>
</dependency>

Configuration​

  • type - string, must be "s3". Required.
  • endpoint - string, the endpoint for S3 compliant service like MinIO, Ceph, etc. Optional
  • bucketName - string, the S3 bucket name. Required
  • fileNamePrefix - string, prefix for the event file names. It is separated from the timestamp with underscore. It can include path and file name prefix. Optional.
Credentials​

To authenticate, the transport uses the default credentials provider chain. The possible authentication methods include:

  • Java system properties
  • Environment variables
  • Shared credentials config file (by default ~/.aws/config)
  • EC2 instance credentials (convenient in EMR and Glue)
  • and other

Refer to the documentation for details.

Behavior​

  • Events are serialized to JSON and stored in the specified S3 bucket.
  • Each event file is named based on its eventTime, converted to epoch milliseconds, with an optional prefix if configured.

Examples​

transport:
  type: s3
  endpoint: https://my-minio.example.com
  bucketName: events
  fileNamePrefix: my/service/events/event

Error Handling via Transport​

// Connect to http://localhost:5000
OpenLineageClient client = OpenLineageClient.builder()
  .transport(
    HttpTransport.builder()
      .uri("http://localhost:5000")
      .apiKey("f38d2189-c603-4b46-bdea-e573a3b5a7d5")
      .build())
  .registerErrorHandler(new EmitErrorHandler() {
    @Override
    public void handleError(Throwable throwable) {
      // Handle emit error here
    }
  }).build();

Defining Your Own Transport​

OpenLineageClient client = OpenLineageClient.builder()
  .transport(
    new MyTransport() {
      @Override
      public void emit(OpenLineage.RunEvent runEvent) {
        // Add emit logic here
      }
    }).build();

Circuit Breakers​

info

This feature is available in OpenLineage versions >= 1.9.0.

To prevent from over-instrumentation OpenLineage integration provides a circuit breaker mechanism that stops OpenLineage from creating, serializing and sending OpenLineage events.

Timeout only Circuit Breaker​

Circuit breaker which closes after a given timeout. It is useful to control the time spent on OpenLineage. Please note that other circuit breakers support timeout as well, but this one is the simplest to fit the scenarios when only timeout is needed.

circuitBreaker:
  type: timeout
  timeoutInSeconds: 90

Simple Memory Circuit Breaker​

This circuit breaker provides a straightforward protective mechanism by monitoring a single metric: the amount of free memory in the JVM. It is a lightweight option ideal for preventing OutOfMemoryError conditions when memory usage is the primary concern.

Triggering Logic

The circuit starts in a closed (operational) state, allowing OpenLineage events to be collected. It will open (trip and temporarily disable OpenLineage) if the percentage of free JVM heap memory drops below the configured memoryThreshold, which is the only condition it checks.

circuitBreaker:
  type: simpleMemory
  memoryThreshold: 20
  circuitCheckIntervalInMillis: 1000
  timeoutInSeconds: 90

Java Runtime Circuit Breaker​

This circuit breaker provides a sophisticated health check by monitoring two key indicators of JVM health: free memory and garbage collection (GC) overhead. It is designed to disable OpenLineage only when the application is both low on memory and actively struggling to reclaim it.

Triggering Logic

The circuit starts in a closed (operational) state. It will open (trip and temporarily disable OpenLineage) only when both of the following conditions are met during a single check:

  1. The percentage of free JVM heap memory drops below the configured memoryThreshold.
  2. The percentage of CPU time spent on Garbage Collection since the last check rises above the configured gcCpuThreshold.

Because both conditions must be true, it allows the application to handle temporary dips in free memory as long as the GC process is not overwhelmed.

Note on Initial State: The GC overhead is calculated as a percentage of time between checks. On the very first check after the application starts, this metric is not yet available. Therefore, the circuit will remain closed (enabled) for the first event, which begins the monitoring cycle.

circuitBreaker:
  type: javaRuntime
  memoryThreshold: 20
  gcCpuThreshold: 10
  circuitCheckIntervalInMillis: 1000
  timeoutInSeconds: 90

Custom Circuit Breaker​

List of available circuit breakers can be extended with custom one loaded via ServiceLoader with own implementation of io.openlineage.client.circuitBreaker.CircuitBreakerBuilder.

Task Queue based Async CircuitBreaker​

High-volume Spark applications can generate an excessive number of events, which can overwhelm the connector and negatively impact the application by choking the shared listener bus.

The TaskQueueCircuitBreaker is designed to mitigate this issue. It manages event processing by adding each task to a bounded queue and handling them asynchronously. To attempt to preserve event order, it waits a configurable amount of time for a task to complete. For critical situations, a close() method allows for abandoning all pending tasks to immediately unblock the listener bus.

circuitBreaker:
  type: asyncTaskQueue
  threadCount: 2
  queueSize: 10
  blockingTimeInSeconds: 1
  shutdownTimeoutSeconds: 60

Metrics​

info

This feature is available in OpenLineage 1.11 and above

To ease the operational experience of using the OpenLineage integrations, this document details the metrics collected by the Java client and the configuration settings for various metric backends.

Metrics collected by Java Client​

The following table outlines the metrics collected by the OpenLineage Java client, which help in monitoring the integration's performance:

MetricDefinitionType
openlineage.emit.startNumber of events the integration started to sendCounter
openlineage.emit.completeNumber of events the integration completed sendingCounter
openlineage.emit.timeTime spent on emitting eventsTimer
openlineage.circuitbreaker.engagedStatus of the Circuit Breaker (engaged or not)Gauge

Metric Backends​

OpenLineage uses Micrometer for metrics collection, similar to how SLF4J operates for logging. Micrometer provides a facade over different metric backends, allowing metrics to be dispatched to various destinations.

Configuring Metric Backends​

Below are the available backends and potential configurations using Micrometer's facilities.

StatsD​

Full configuration options for StatsD can be found in the Micrometer's StatsDConfig implementation.

metrics:
  type: statsd
  flavor: datadog
  host: localhost
  port: 8125

Dataset Namespace Resolver​

info

This feature is available in OpenLineage 1.17 and above

Oftentimes host addresses are used to access data and a single dataset can be accessed via different addresses. For example, a Kafka topic can be accessed by a list of kafka bootstrap servers or any server from the list. In general, a problem can be solved by adding mechanism which resolves host addresses into logical identifier understood within the organisation. This applies for all clusters like Kafka or Cassandra which should be identified regardless of current list of hosts they contain. This also applies for JDBC urls where a physical address of database can change over time.

Host List Resolver​

Host List Resolver given a list of hosts, replaces host name within the dataset namespace into the resolved value defined.

dataset:
  namespaceResolvers:
    resolved-name:
      type: hostList
      hosts: ['kafka-prod13.company.com', 'kafka-prod15.company.com']
      schema: "kafka"

Pattern Namespace Resolver​

Java regex pattern is used to identify a host. Substrings matching a pattern will be replaced with resolved name.

dataset:
  namespaceResolvers:
    resolved-name:
      type: pattern
      # 'cassandra-prod7.company.com', 'cassandra-prod8.company.com'
      regex: 'cassandra-prod(\d)+\.company\.com'
      schema: "cassandra"

Pattern Group Namespace Resolver​

For this resolver, Java regex pattern is used to identify a host. However, instead of configured resolved name, a matchingGroup is used a resolved name. This can be useful when having several clusters made from hosts with a well-defined host naming convention.

dataset:
  namespaceResolvers:
    test-pattern:
      type: patternGroup
      # 'cassandra-test-7.company.com', 'cassandra-test-8.company.com', 'kafka-test-7.company.com', 'kafka-test-8.company.com'
      regex: '(?<cluster>[a-zA-Z-]+)-(\d)+\.company\.com:[\d]*'
      matchingGroup: "cluster"
      schema: "cassandra"

Custom Resolver​

Custom resolver can be added by implementing:

  • io.openlineage.client.dataset.namespaceResolver.DatasetNamespaceResolver
  • io.openlineage.client.dataset.namespaceResolver.DatasetNamespaceResolverBuilder
  • io.openlineage.client.dataset.namespaceResolver.DatasetNamespaceResolverConfig

Config class can be used to pass any namespace resolver parameters through standard configuration mechanism (Spark & Flink configuration or openlineage.yml file provided). Standard ServiceLoader approach is used to load and initiate custom classes.

Tags​

You can specify tags for Runs and Jobs using the following syntax: Tags added by run will be added as TagsRunFacet, while tags added by jobs will be added as TagsJobFacet.

run:
    tags:
      - "key:value"
      - "label"
      - "key:value:source"
jobs:
    tags: [
      "key:value",
      "label",
      "key:value:source"
    ]

Dataset Name Normalization​

Sometimes, an object storage path used by a job to read or write data does not represent a proper dataset name.
To address this, a dataset name trimmer can be applied to trim trailing name segments that are not part of the actual dataset name.

How It Works​

  • The trimmed dataset name becomes the dataset name.
  • The full, non-trimmed dataset name is stored in the subset definition facet as a LocationSubsetCondition.

Why It Matters​

This approach is especially useful for input datasets, where multiple paths may point to the same directory.

  • The subset definition facet captures all directories read.
  • This reduces the size of OpenLineage events by avoiding duplication, since otherwise each directory would be treated as a separate dataset.

Reducing Datasets in Java Client​

Datasets are reduced only if:

  1. Their names are trimmed to the same dataset name.
  2. They share identical facets.

By the default, OpenLineage Java client comes with the following trimmers:

  • io.openlineage.client.dataset.partition.trimmer.DateTrimmer
  • io.openlineage.client.dataset.partition.trimmer.KeyValueTrimmer
  • io.openlineage.client.dataset.partition.trimmer.MultiDirTrimmer
  • io.openlineage.client.dataset.partition.trimmer.YearMonthTrimmer

The list of the trimmers can be managed by disabledTrimmers and extraTrimmers configuration parameters.

In most cases, trimmers work on the last directory segment of the dataset name. The trimming process runs iteratively, applying trimmers repeatedly until no additional segments can be removed.

Trimmers Configuration​

dataset:
  disabledTrimmers: io.openlineage.client.dataset.partition.trimmer.DateTrimmer
  extraTrimmers: org.company.CustomTrimmer

Out of the box trimmers​

DateTrimmer​

Remove a trailing date partition. It checks if the last part of the dataset name contains a valid and recognized date pattern. Then it checks if the other characters in the directory are only numeric and non-numeric T and Z characters. This behaviour assures agility to detect dates beyond the common formats configured in the trimmer.

  • .../20250901/ → trims /20250901/
  • .../2025-09-01/ → trims /2025-09-01/
  • .../20250722T901Z/ → trims /20250722T901Z/ as it contains a valid date pattern with extra digits and non-numeric T and Z characters only.
  • .../2025-25-01/ → trims nothing as it is not a valid date
  • .../dt=2025-09-01/ → may be handled by KeyValueTrimmer

KeyValueTrimmer​

Remove last part of the dataset name if it follows key=value pattern.

  • .../dt=2025-09-01/ → trims /dt=2025-09-01/
  • .../hour=05/ → trims /hour=05/

MultiDirDateTrimmer​

Trims multiple directories at once if they are valid date or year month.

  • .../2025/09/01/ → trims /2025/09/01/
  • .../2025/09/ → trims /2025/09/

YearMonthTrimmer​

Trims trailing directory if it is a valid year and month.

  • .../202509/ → trims /2025/09/
  • .../202533/ → trims nothing
  • .../2025-09/ → trims /2025-09/