Environment Variables

The following is a summary of a few important environment variables which expose various levers which control how DataHub works.

Feature Flags

Variable	Default	Unit/Type	Components	Description
UI_INGESTION_ENABLED	true	boolean	[GMS, MCE Consumer]	Enable UI based ingestion.
DATAHUB_ANALYTICS_ENABLED	true	boolean	[Frontend, GMS]	Collect DataHub usage to populate the analytics dashboard.
BOOTSTRAP_SYSTEM_UPDATE_WAIT_FOR_SYSTEM_UPDATE	true	boolean	[GMS, MCE Consumer, MAE Consumer]	Do not wait for the system-update to complete before starting. This should typically only be disabled during development.
ER_MODEL_RELATIONSHIP_FEATURE_ENABLED	false	boolean	[Frontend, GMS]	Enable ER Model Relation Feature that shows Relationships Tab within a Dataset UI.
STRICT_URN_VALIDATION_ENABLED	false	boolean	[GMS, MCE Consumer, MAE Consumer]	Enable stricter URN validation logic

Ingestion

Variable	Default	Unit/Type	Components	Description
ASYNC_INGEST_DEFAULT	false	boolean	[GMS]	Asynchronously process ingestProposals by writing the ingestion MCP to Kafka. Typically enabled with standalone consumers.
MCP_CONSUMER_ENABLED	true	boolean	[GMS, MCE Consumer]	When running in standalone mode, disabled on GMS and enabled on separate MCE Consumer.
MCL_CONSUMER_ENABLED	true	boolean	[GMS, MAE Consumer]	When running in standalone mode, disabled on GMS and enabled on separate MAE Consumer.
PE_CONSUMER_ENABLED	true	boolean	[GMS, MAE Consumer]	When running in standalone mode, disabled on GMS and enabled on separate MAE Consumer.
ES_BULK_REQUESTS_LIMIT	1000	docs	[GMS, MAE Consumer]	Number of bulk documents to index. MAE Consumer if standalone.
ES_BULK_FLUSH_PERIOD	1	seconds	[GMS, MAE Consumer]	How frequently indexed documents are made available for query.
ALWAYS_EMIT_CHANGE_LOG	false	boolean	[GMS]	Enables always emitting a MCL even when no changes are detected. Used for Time Based Lineage when no changes occur.
GRAPH_SERVICE_DIFF_MODE_ENABLED	true	boolean	[GMS]	Enables diff mode for graph writes, uses a different code path that produces a diff from previous to next to write relationships instead of wholesale deleting edges and reading.

Caching

Variable	Default	Unit/Type	Components	Description
SEARCH_SERVICE_ENABLE_CACHE	false	boolean	[GMS]	Enable caching of search results.
SEARCH_SERVICE_CACHE_IMPLEMENTATION	caffeine	string	[GMS]	Set to hazelcast if the number of GMS replicas > 1 for enabling distributed cache.
CACHE_TTL_SECONDS	600	seconds	[GMS]	Default cache time to live.
CACHE_MAX_SIZE	10000	objects	[GMS]	Maximum number of items to cache.
LINEAGE_SEARCH_CACHE_ENABLED	true	boolean	[GMS]	Enables in-memory cache for searchAcrossLineage query.
CACHE_ENTITY_COUNTS_TTL_SECONDS	600	seconds	[GMS]	Homepage entity count time to live.
CACHE_SEARCH_LINEAGE_TTL_SECONDS	86400	seconds	[GMS]	Search lineage cache time to live.
CACHE_SEARCH_LINEAGE_LIGHTNING_THRESHOLD	300	objects	[GMS]	Lineage graphs exceeding this limit will use a local cache.

Search

Variable	Default	Unit/Type	Components	Description
INDEX_PREFIX	``	string	[GMS, MAE Consumer, Elasticsearch Setup, System Update]	Prefix Elasticsearch indices with the given string.
ELASTICSEARCH_NUM_SHARDS_PER_INDEX	1	integer	[System Update]	Default number of shards per Elasticsearch index.
ELASTICSEARCH_NUM_REPLICAS_PER_INDEX	1	integer	[System Update]	Default number of replica per Elasticsearch index.
ELASTICSEARCH_BUILD_INDICES_RETENTION_VALUE	60	integer	[System Update]	Number of units for the retention of Elasticsearch clone/backup indices.
ELASTICSEARCH_BUILD_INDICES_RETENTION_UNIT	DAYS	string	[System Update]	Unit for the retention of Elasticsearch clone/backup indices.
ELASTICSEARCH_QUERY_EXACT_MATCH_EXCLUSIVE	false	boolean	[GMS]	Only return exact matches when using quotes.
ELASTICSEARCH_QUERY_EXACT_MATCH_WITH_PREFIX	true	boolean	[GMS]	Include prefix match in exact match results.
ELASTICSEARCH_QUERY_EXACT_MATCH_FACTOR	10.0	float	[GMS]	Multiply by this number on true exact match.
ELASTICSEARCH_QUERY_EXACT_MATCH_PREFIX_FACTOR	1.6	float	[GMS]	Multiply by this number when prefix match.
ELASTICSEARCH_QUERY_EXACT_MATCH_CASE_FACTOR	0.7	float	[GMS]	Multiply by this number when case insensitive match.
ELASTICSEARCH_QUERY_EXACT_MATCH_ENABLE_STRUCTURED	true	boolean	[GMS]	When using structured query, also include exact matches.
ELASTICSEARCH_QUERY_PARTIAL_URN_FACTOR	0.5	float	[GMS]	Multiply by this number when partial token match on URN)
ELASTICSEARCH_QUERY_PARTIAL_FACTOR	0.4	float	[GMS]	Multiply by this number when partial token match on non-URN field.
ELASTICSEARCH_QUERY_CUSTOM_CONFIG_ENABLED	true	boolean	[GMS]	Enable search query and ranking customization configuration.
ELASTICSEARCH_QUERY_CUSTOM_CONFIG_FILE	search_config.yml	string	[GMS]	The location of the search customization configuration.
ELASTICSEARCH_INDEX_BUILDER_MAPPINGS_REINDEX	false	boolean	[System Update]	Enable reindexing on Elasticsearch schema changes.
ENABLE_STRUCTURED_PROPERTIES_SYSTEM_UPDATE	false	boolean	[System Update]	Enable reindexing to remove hard deleted structured properties.

Kafka

In general, there are lots of Kafka configuration environment variables for both the producer and consumers defined in the official Spring Kafka documentation here. These environment variables follow the standard Spring representation of properties as environment variables. Simply replace the dot, ., with an underscore, _, and convert to uppercase.

Variable	Default	Unit/Type	Components	Description
KAFKA_LISTENER_CONCURRENCY	1	integer	[GMS, MCE Consumer, MAE Consumer]	Number of Kafka consumer threads. Optimize throughput by matching to topic partitions.
SPRING_KAFKA_PRODUCER_PROPERTIES_MAX_REQUEST_SIZE	1048576	bytes	[GMS, MCE Consumer, MAE Consumer]	Max produced message size. Note that the topic configuration is not controlled by this variable.
SCHEMA_REGISTRY_TYPE	INTERNAL	string	[GMS, MCE Consumer, MAE Consumer]	Schema registry implementation. One of INTERNAL or KAFKA or AWS_GLUE
KAFKA_SCHEMAREGISTRY_URL	http://localhost:8080/schema-registry/api/	string	[GMS, MCE Consumer, MAE Consumer]	Schema registry url. Used for INTERNAL and KAFKA. The default value is for the GMS component. The MCE Consumer and MAE Consumer should be the GMS hostname and port.
AWS_GLUE_SCHEMA_REGISTRY_REGION	us-east-1	string	[GMS, MCE Consumer, MAE Consumer]	If using AWS_GLUE in the SCHEMA_REGISTRY_TYPE variable for the schema registry implementation.
AWS_GLUE_SCHEMA_REGISTRY_NAME	``	string	[GMS, MCE Consumer, MAE Consumer]	If using AWS_GLUE in the SCHEMA_REGISTRY_TYPE variable for the schema registry.
USE_CONFLUENT_SCHEMA_REGISTRY	true	boolean	[kafka-setup]	Enable Confluent schema registry configuration.
KAFKA_PRODUCER_MAX_REQUEST_SIZE	5242880	integer	[Frontend, GMS, MCE Consumer, MAE Consumer]	Max produced message size. Note that the topic configuration is not controlled by this variable.
KAFKA_CONSUMER_MAX_PARTITION_FETCH_BYTES	5242880	integer	[GMS, MCE Consumer, MAE Consumer]	The maximum amount of data per-partition the server will return. Records are fetched in batches by the consumer. If the first record batch in the first non-empty partition of the fetch is larger than this limit, the batch will still be returned to ensure that the consumer can make progress.
MAX_MESSAGE_BYTES	5242880	integer	[kafka-setup]	Sets the max message size on the kakfa topics.
KAFKA_PRODUCER_COMPRESSION_TYPE	snappy	string	[Frontend, GMS, MCE Consumer, MAE Consumer]	The compression used by the producer.

Frontend

Variable	Default	Unit/Type	Components	Description
AUTH_VERBOSE_LOGGING	false	boolean	[Frontend]	Enable verbose authentication logging. Enabling this will leak sensisitve information in the logs. Disable when finished debugging.
AUTH_OIDC_GROUPS_CLAIM	groups	string	[Frontend]	Claim to use as the user's group.
AUTH_OIDC_EXTRACT_GROUPS_ENABLED	false	boolean	[Frontend]	Auto-provision the group from the user's group claim.
AUTH_SESSION_TTL_HOURS	24	string	[Frontend]	The number of hours a user session is valid. After this many hours the actor cookie will be expired by the browser and the user will be prompted to login again.
MAX_SESSION_TOKEN_AGE	24h	string	[Frontend]	The maximum age of the session token. User session tokens are stateless and will become invalid after this time requiring a user to login again.