feat(ingest/cockroachdb): add cockroachdb ingestion

datahub-web-react/src/app/ingest/source/builder/constants.ts

@@ -16,6 +16,7 @@ import supersetLogo from '../../../../images/supersetlogo.png';
 import athenaLogo from '../../../../images/awsathenalogo.png';
 import mssqlLogo from '../../../../images/mssqllogo.png';
 import clickhouseLogo from '../../../../images/clickhouselogo.png';
+import cockroachdbLogo from '../../../../images/cockroachdblogo.png';
 import trinoLogo from '../../../../images/trinologo.png';
 import dbtLogo from '../../../../images/dbtlogo.png';
 import druidLogo from '../../../../images/druidlogo.png';
@@ -44,6 +45,8 @@ export const BIGQUERY_URN = `urn:li:dataPlatform:${BIGQUERY}`;
 export const CLICKHOUSE = 'clickhouse';
 export const CLICKHOUSE_USAGE = 'clickhouse-usage';
 export const CLICKHOUSE_URN = `urn:li:dataPlatform:${CLICKHOUSE}`;
+export const COCKROACHDB = 'cockroachdb';
+export const COCKROACHDB_URN = `urn:li:dataPlatform:${COCKROACHDB}`;
 export const DBT = 'dbt';
 export const DBT_URN = `urn:li:dataPlatform:${DBT}`;
 export const DRUID = 'druid';
@@ -122,6 +125,7 @@ export const PLATFORM_URN_TO_LOGO = {
     [AZURE_URN]: azureLogo,
     [BIGQUERY_URN]: bigqueryLogo,
     [CLICKHOUSE_URN]: clickhouseLogo,
+    [COCKROACHDB_URN]: cockroachdbLogo,
     [DBT_URN]: dbtLogo,
     [DRUID_URN]: druidLogo,
     [DYNAMODB_URN]: dynamodbLogo,

datahub-web-react/src/app/ingest/source/builder/sources.json

@@ -243,5 +243,12 @@
         "displayName": "Qlik Sense",
         "docsUrl": "https://datahubproject.io/docs/generated/ingestion/sources/qlik-sense/",
         "recipe": "source:\n    type: qlik-sense\n    config:\n        # Coordinates\n        tenant_hostname: https://xyz12xz.us.qlikcloud.com\n        # Coordinates\n        api_key: QLIK_API_KEY\n\n        # Optional - filter for certain space names instead of ingesting everything.\n        # space_pattern:\n\n        #   allow:\n        #     - space_name\n        ingest_owner: true"
+    },
+    {
+        "urn": "urn:li:dataPlatform:cockroachdb",
+        "name": "cockroachdb",
+        "displayName": "CockroachDb",
+        "docsUrl": "https://datahubproject.io/docs/generated/ingestion/sources/cockroachdb/",
+        "recipe": "source: \n    type: cockroachdb\n    config:\n        # Coordinates\n        host_port: # Your CockroachDb host and port, e.g. cockroachdb:5432\n        database: # Your CockroachDb Database, e.g. sample_db\n\n        # Credentials\n        # Add secret in Secrets Tab with relevant names for each variable\n        username: null # Your CockroachDb username, e.g. admin\n\n        # Options\n        include_tables: true\n        include_views: true\n\n        # Profiling\n        profiling:\n            enabled: true\n            profile_table_level_only: true\n        stateful_ingestion:\n            enabled: true"
     }
 ]

metadata-ingestion/docs/sources/cockroachdb/cockroachdb_recipe.yml

@@ -0,0 +1,21 @@
+source:
+  type: cockroachdb
+  config:
+    # Coordinates
+    host_port: localhost:26257
+    database: DemoDatabase
+
+    # Credentials
+    username: user
+    password: pass
+
+    # Optional: SSL configuration.
+    # options:
+    #   connect_args:
+    #     sslcert: "<<path to sslcert>>"
+    #     sslkey: "<<path to sslkey>>"
+    #     sslrootcert: "<<path to verification ca chain>>"
+    #     sslmode: "verify-full"
+
+sink:
+  # sink configs

metadata-ingestion/setup.py

@@ -311,6 +311,7 @@
     | classification_lib,
     "clickhouse": sql_common | clickhouse_common,
     "clickhouse-usage": sql_common | usage_common | clickhouse_common,
+    "cockroachdb": sql_common | {"psycopg2-binary", "GeoAlchemy2", "sqlalchemy-cockroachdb<2.0.0"},
     "datahub-lineage-file": set(),
     "datahub-business-glossary": set(),
     "delta-lake": {*data_lake_profiling, *delta_lake},
@@ -505,6 +506,7 @@
             "bigquery",
             "clickhouse",
             "clickhouse-usage",
+            "cockroachdb",
             "delta-lake",
             "druid",
             "elasticsearch",
@@ -596,6 +598,7 @@
         "bigquery = datahub.ingestion.source.bigquery_v2.bigquery:BigqueryV2Source",
         "clickhouse = datahub.ingestion.source.sql.clickhouse:ClickHouseSource",
         "clickhouse-usage = datahub.ingestion.source.usage.clickhouse_usage:ClickHouseUsageSource",
+        "cockroachdb = datahub.ingestion.source.sql.cockroachdb:CockroachDBSource",
         "delta-lake = datahub.ingestion.source.delta_lake:DeltaLakeSource",
         "s3 = datahub.ingestion.source.s3:S3Source",
         "dbt = datahub.ingestion.source.dbt.dbt_core:DBTCoreSource",

metadata-ingestion/src/datahub/ingestion/source/sql/cockroachdb.py

@@ -0,0 +1,45 @@
+from pydantic.fields import Field
+
+from datahub.configuration.common import AllowDenyPattern
+from datahub.ingestion.api.common import PipelineContext
+from datahub.ingestion.api.decorators import (
+    SourceCapability,
+    SupportStatus,
+    capability,
+    config_class,
+    platform_name,
+    support_status,
+)
+from datahub.ingestion.source.sql.postgres import PostgresConfig, PostgresSource
+
+
+class CockroachDBConfig(PostgresConfig):
+    scheme = Field(default="cockroachdb+psycopg2", description="database scheme")
+    schema_pattern = Field(
+        default=AllowDenyPattern(deny=["information_schema", "crdb_internal"])
+    )
+
+
+@platform_name("CockroachDB")
+@config_class(CockroachDBConfig)
+@support_status(SupportStatus.TESTING)
+@capability(SourceCapability.PLATFORM_INSTANCE, "Enabled by default")
+@capability(SourceCapability.DOMAINS, "Supported via the `domain` config field")
+@capability(SourceCapability.DATA_PROFILING, "Optionally enabled via configuration")
+@capability(SourceCapability.DELETION_DETECTION, "Enabled via stateful ingestion")
+class CockroachDBSource(PostgresSource):
+
+    config: CockroachDBConfig
+
+    def __init__(self, config: CockroachDBConfig, ctx: PipelineContext):
+        super().__init__(config, ctx)
+
+    def get_platform(self):
+        return "cockroachdb"
+
+    @classmethod
+    def create(cls, config_dict, ctx):
+        config = CockroachDBConfig.parse_obj(config_dict)
+        return cls(config, ctx)
+
+    pass

metadata-ingestion/src/datahub/ingestion/source/sql/postgres.py

@@ -145,7 +145,10 @@ class PostgresSource(SQLAlchemySource):
     config: PostgresConfig
 
     def __init__(self, config: PostgresConfig, ctx: PipelineContext):
-        super().__init__(config, ctx, "postgres")
+        super().__init__(config, ctx, self.get_platform())
+
+    def get_platform(self):
+        return "postgres"
 
     @classmethod
     def create(cls, config_dict, ctx):

metadata-ingestion/tests/unit/test_cockroach_source.py

@@ -0,0 +1,26 @@
+from datahub.ingestion.api.common import PipelineContext
+from datahub.ingestion.source.sql.cockroachdb import (
+    CockroachDBConfig,
+    CockroachDBSource,
+)
+from datahub.ingestion.source.sql.postgres import PostgresConfig, PostgresSource
+
+
+def _base_config():
+    return {"username": "user", "password": "password", "host_port": "host:1521"}
+
+
+def test_platform_correctly_set_cockroachdb():
+    source = CockroachDBSource(
+        ctx=PipelineContext(run_id="cockroachdb-source-test"),
+        config=CockroachDBConfig.parse_obj(_base_config()),
+    )
+    assert source.platform == "cockroachdb"
+
+
+def test_platform_correctly_set_postgres():
+    source = PostgresSource(
+        ctx=PipelineContext(run_id="postgres-source-test"),
+        config=PostgresConfig.parse_obj(_base_config()),
+    )
+    assert source.platform == "postgres"

metadata-service/war/src/main/resources/boot/data_platforms.json

@@ -48,6 +48,16 @@
       "logoUrl": "/assets/platforms/clickhouselogo.png"
     }
   },
+  {
+    "urn": "urn:li:dataPlatform:cockroachdb",
+    "aspect": {
+      "datasetNameDelimiter": ".",
+      "name": "cockroachdb",
+      "displayName": "CockroachDb",
+      "type": "RELATIONAL_DB",
+      "logoUrl": "/assets/platforms/cockroachdblogo.png"
+    }
+  },
   {
     "urn": "urn:li:dataPlatform:couchbase",
     "aspect": {