Merge branch 'datahub-project:master' into master

acryldata · May 24, 2024 · 2c9d395 · 2c9d395
2 parents d9d4c96 + 47ba6d9
commit 2c9d395
Show file tree

Hide file tree

Showing 21 changed files with 648 additions and 356 deletions.
diff --git a/docs-website/sidebars.js b/docs-website/sidebars.js
@@ -243,7 +243,7 @@ module.exports = {
           "Operator Guide": [
             {
               type: "doc",
-              id: "docs/managed-datahub/operator-guide/setting-up-remote-ingestion-executor-on-aws",
+              id: "docs/managed-datahub/operator-guide/setting-up-remote-ingestion-executor",
               className: "saasOnly",
             },
             {

diff --git a/docs/managed-datahub/managed-datahub-overview.md b/docs/managed-datahub/managed-datahub-overview.md
@@ -109,8 +109,8 @@ Fill out
 ## Additional Integrations
 
 - [Slack Integration](docs/managed-datahub/saas-slack-setup.md)
+- [Remote Ingestion Executor](docs/managed-datahub/operator-guide/setting-up-remote-ingestion-executor.md)
 - [AWS Privatelink](docs/managed-datahub/integrations/aws-privatelink.md)
-- [AWS Ingestion Executor](docs/managed-datahub/operator-guide/setting-up-remote-ingestion-executor-on-aws.md)
 - [AWS Eventbridge](docs/managed-datahub/operator-guide/setting-up-events-api-on-aws-eventbridge.md)
 
 ## Additional SSO/Login Support

diff --git a/...ng-up-remote-ingestion-executor-on-aws.md → ...e/setting-up-remote-ingestion-executor.md b/...ng-up-remote-ingestion-executor-on-aws.md → ...e/setting-up-remote-ingestion-executor.md
@@ -6,48 +6,50 @@ description: >-
 ---
 import FeatureAvailability from '@site/src/components/FeatureAvailability';
 
-# Setting up Remote Executor on AWS
+# Setting up Remote Ingestion Executor
 <FeatureAvailability saasOnly />
 
 ## Overview
 
-> [!NOTE]
-> Acryl Remote Executor can now be used to both run ingestions from and monitor ingestion sources that are not publicly accessible via the internet.
+:::note
 
-Acryl DataHub comes packaged with an Acryl-managed executor, which is hosted inside of Acryl's environment on your behalf. However, there are certain scenarios in which an Acryl-hosted executor is not sufficient to cover all of an organization's ingestion sources.
+The Remote Executor can now also be used for monitoring of ingestion sources.
 
-For example, if an ingestion source is not publicly accessible via the internet, e.g. hosted privately within a specific AWS account, then the Acryl executor will be unable to extract metadata from it.
+:::
 
+Acryl DataHub comes packaged with an Acryl-managed executor, which is hosted inside of Acryl's environment on your behalf. However, there are certain scenarios in which an Acryl-hosted executor is not ideal or sufficient to cover all of an organization's ingestion sources. For example, if an ingestion source is hosted behind a firewall or in an environment with strict access policies, then the Acryl executor might be unable to connect to it to extract metadata.
 
-<p align="center">
-  <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/image-(12).png"/>
-</p>
+To accommodate these cases, Acryl supports configuring a Remote Ingestion Executor which can be deployed inside of your environment – whether that is on-prem or in cloud. This setup allows you to continue leveraging the Acryl DataHub console to create, schedule, and run both ingestion and assertion monitors, all while retaining network and credential isolation.
 
+## Deploying a Remote Executor
 
-To accommodate these cases, Acryl supports configuring a remote executor which can be deployed inside of your AWS account. This setup allows you to continue leveraging the Acryl DataHub console to create, schedule, and run both ingestion and assertion monitors, all while retaining network and credential isolation.
+:::note
 
+The Remote Ingestion Executor is only available for Managed DataHub. Setting up a new executor requires coordination with an Acryl representative.
 
-<p align="center">
-  <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/image-(6).png"/>
-</p>
+:::
 
+The Remote Ingestion Executor can be deployed on several different platforms, including [Amazon ECS](#deploying-on-amazon-ecs), [Kubernetes](#deploying-on-kubernetes) (GKE, EKS or self-hosted) and others. It can also be deployed using several different methods, including [CloudFormation](https://raw.githubusercontent.com/acryldata/datahub-cloudformation/master/remote-executor/datahub-executor.ecs.template.yaml) or [Terraform](https://github.com/acryldata/datahub-terraform-modules/tree/main/remote-ingestion-executor) templates for ECS, or [Helm chart](https://github.com/acryldata/datahub-executor-helm) for Kubernetes. Please reach out to your Acryl representative for other alternatives.
 
-## Deploying a Remote Executor
 
-> [!NOTE]
-> Customers migrating from the legacy DataHub Executor: migration to the new executor requires a configuration change on Acryl side. Please contact your Acryl representative for detailed guidance.
->
-> Steps you will need to perform on your end when instructed by your Acryl representative:
-> 1. Temporarily stop your legacy DataHub Remote Executor instance (e.g. `aws ecs update-service --desired-count 0 --cluster "cluster-name" --service "service-name"`)
-> 2. Deploy new DataHub Executor using steps below.
-> 3. Trigger an ingestion to make sure the new executor is working as expected.
-> 4. Tear down legacy executor ECR deployment.
+### Deploying on Amazon ECS
 
+:::note
 
-1. **Provide AWS Account Id**: Provide Acryl Team with the id of the AWS in which the remote executor will be hosted. This will be used to grant access to private Acryl ECR registry. The account id can be provided to your Acryl representative via Email or [One Time Secret](https://onetimesecret.com/).
+Customers migrating from the legacy DataHub Executor: migration to the new executor requires a configuration change on Acryl side. Please contact your Acryl representative for detailed guidance.
 
-2. **Provision an Acryl Executor** (ECS)**:** Acryl team will provide a [Cloudformation Template](https://raw.githubusercontent.com/acryldata/datahub-cloudformation/master/remote-executor/datahub-executor.ecs.template.yaml) that you can run to provision an ECS cluster with a single remote ingestion task. It will also provision an AWS role for the task which grants the permissions necessary to read and delete from the private SQS queue created for you, along with reading the secrets you've specified. At minimum, the template requires the following parameters:
-   1. **Deployment Location:** The AWS VPC + subnet in which the Acryl Executor task is to be provisioned.
+Steps you will need to perform on your end when instructed by your Acryl representative:
+1. Temporarily stop your legacy DataHub Remote Executor instance (e.g. `aws ecs update-service --desired-count 0 --cluster "cluster-name" --service "service-name"`)
+2. Deploy new DataHub Executor using steps below.
+3. Trigger an ingestion to make sure the new executor is working as expected.
+4. Tear down legacy executor ECS deployment.
+
+:::
+
+1. **Provide AWS account ID**: Provide Acryl with the ID of the AWS account in which the remote executor will be hosted. This will be used to grant access to the private Acryl ECR registry. The account ID can be provided to your Acryl representative via Email or [One Time Secret](https://onetimesecret.com/).
+
+2. **Provision an Acryl Executor** (ECS)**:** Acryl team will provide a [Cloudformation Template](https://raw.githubusercontent.com/acryldata/datahub-cloudformation/master/remote-executor/datahub-executor.ecs.template.yaml) that you can run to provision an ECS cluster with a single remote ingestion task. It will also provision an AWS role for the task which grants the permissions necessary to read and delete from the private queue created for you, along with reading the secrets you've specified. At minimum, the template requires the following parameters:
+   1. **Deployment Location:** The AWS VPC and subnet in which the Acryl Executor task is to be provisioned.
    2. **DataHub Personal Access Token**: A valid DataHub PAT. This can be generated inside of **Settings > Access Tokens** of DataHub web application. You can alternatively create a secret in AWS Secrets Manager and refer to that by ARN.
    3. **Acryl DataHub URL**: The URL for your DataHub instance, e.g. `<your-company>.acryl.io/gms`. Note that you MUST enter the trailing /gms when configuring the executor.
    4. **Acryl Remote Executor Version:** The version of the remote executor to deploy. This is converted into a container image tag. It will be set to the latest version of the executor by default.
@@ -58,42 +60,12 @@ To accommodate these cases, Acryl supports configuring a remote executor which c
 
        Note that the only external secret provider that is currently supported is AWS Secrets Manager.
 
-   **Alternatively**: if you prefer to deploy Acryl Remote Executor on your private Kubernetes cluster, please use [DataHub Executor Helm chart](https://github.com/acryldata/datahub-executor-helm/tree/main/charts/datahub-executor-worker). Similar to the above, you would need to provide the following parameters when installing the chart:
-   1. **Create a secret with DataHub Access Token**. A valid DataHub PAT. You can create the secret object as follows:
-        ```
-        $ kubectl create secret generic datahub-access-token-secret --from-literal=datahub-access-token-secret-key=<DATAHUB-ACCESS-TOKEN>
-        ```
-   2. **Acryl DataHub URL**: The URL for your DataHub instance, e.g. `<your-company>.acryl.io/gms`. Note that you MUST enter the trailing /gms when configuring the executor.
-   3. **Acryl Remote Executor Version:** The version of the remote executor to deploy. This is converted into a container image tag.
-   4. **Optionally** pass source secrets/environment variables as necessary.
-
-   Install DataHub Executor chart as follows:
-
-   ```
-   $ helm install \
-     --set global.datahub.executor.worker_id="remote" \
-     --set global.datahub.gms.url="https://company.acryl.io/gms" \
-     --set image.tag=v0.3.1 \
-       default datahub-executor-worker
-   ```
-
-<p align="center">
-  <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/Screen-Shot-2023-01-19-at-5.12.47-PM.png"/>
-</p>
-
-
-
-<p align="center">
-  <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/Screen-Shot-2023-01-19-at-5.12.56-PM.png"/>
-</p>
-
-
 3.  **Test the Executor:** To test your remote executor:
 
-    1. Create a new Ingestion Source by clicking '**Create new Source**' the '**Ingestion**' tab of the DataHub console. Configure your Ingestion Recipe as though you were running it from inside of your environment.
+    1. Create a new Ingestion Source by clicking **Create new Source** in the **Ingestion** tab of the DataHub console. Configure your Ingestion Recipe as though you were running it from inside of your environment.
     2.  When working with "secret" fields (passwords, keys, etc), you can refer to any "self-managed" secrets by name: `${SECRET_NAME}:`
 
-        
+
 <p align="center">
   <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/Screen-Shot-2023-01-19-at-4.16.52-PM.png"/>
 </p>
@@ -109,9 +81,9 @@ To accommodate these cases, Acryl supports configuring a remote executor which c
   <img width="70%"  src="https://raw.githubusercontent.com/datahub-project/static-assets/main/imgs/saas/Screen-Shot-2022-03-07-at-10.23.31-AM.png"/>
 </p>
 
-## Updating a Remote Executor
+#### Updating a deployment
 In order to update the executor, ie. to deploy a new container version, you'll need to update the CloudFormation Stack to re-deploy the CloudFormation template with a new set of parameters.
-### Steps - AWS Console
+##### Steps - AWS Console
 1. Navigate to CloudFormation in AWS Console
 2. Select the stack dedicated to the remote executor
 3. Click **Update**
@@ -128,17 +100,43 @@ In order to update the executor, ie. to deploy a new container version, you'll n
 9. Click **Next**
 10. Confirm your parameter changes, and update. This should perform the necessary upgrades.
 
+### Deploying on Kubernetes
+
+The Helm chart [datahub-executor-worker](https://github.com/acryldata/datahub-executor-helm/tree/main/charts/datahub-executor-worker) can be used to deploy on a Kubernetes cluster. These instructions also apply for deploying to Amazon Elastic Kubernetes Service (EKS) or Google Kubernetes Engine (GKE).
+
+1. **Download Chart**: Download the [latest release](https://github.com/acryldata/datahub-executor-helm/releases) of the chart
+2. **Unpack the release archive**:
+  ```
+  tar zxvf v0.0.4.tar.gz --strip-components=2
+  ```
+3. **Contact Acryl representative**: To provision the required access and infrastructure. The remote executor image is hosted on a private registry. For access within AWS, you will need to provide the IAM principal which will be allowed to pull from the ECR repository. For Google Cloud, you will need to provide the cluster's IAM service account.
+4. **Create a secret with DataHub PAT**: This can be generated under **Settings > Access Tokens** on DataHub. You can create the secret object as follows:
+  ```
+  kubectl create secret generic datahub-access-token-secret --from-literal=datahub-access-token-secret-key=<DATAHUB-ACCESS-TOKEN>
+  ```
+5. **Acryl DataHub URL**: The URL for your DataHub instance, e.g. `https://<your-company>.acryl.io/gms`. Note that you MUST enter the trailing /gms when configuring the executor.
+6. **Acryl Remote Executor Version:** The version of the remote executor to deploy. This is converted into a container image tag.
+7. **Optionally** pass source secrets/environment variables as necessary. See [values.yaml](https://github.com/acryldata/datahub-executor-helm/blob/main/charts/datahub-executor-worker/values.yaml) file for all the available configurable options.
+8. **Install DataHub Executor chart as follows:**
+  ```
+    helm install \
+    --set global.datahub.executor.worker_id="remote" \
+    --set global.datahub.gms.url="https://company.acryl.io/gms" \
+    --set image.tag=v0.3.1 \
+    acryl datahub-executor-worker
+  ```
+
 ## FAQ
 
-### If I need to change (or add) a secret that is stored in AWS Secrets Manager, e.g. for rotation, will the new secret automatically get picked up by Acryl's executor?**
+### If I need to change (or add) a secret that is stored in AWS Secrets Manager, e.g. for rotation, will the new secret automatically get picked up by Acryl's executor?
 
 Unfortunately, no. Secrets are wired into the executor container at deployment time, via environment variables. Therefore, the ECS Task will need to be restarted (either manually or via a stack parameter update) whenever your secrets change.
 
-### I want to deploy multiple Acryl Executors. Is this currently possible?**
+### I want to deploy multiple Acryl Executors. Is this currently possible?
 
-This is possible, but currently requires a configuration change on Acryl side. Please contact your Acryl representative for more information.
+Yes. Please contact your Acryl representative for details.
 
-### I've run the CloudFormation Template, how can I tell that the container was successfully deployed?**
+### I've run the CloudFormation Template, how can I tell that the container was successfully deployed?
 
 We recommend verifying in AWS Console by navigating to **ECS > Cluster > Stack Name > Services > Logs.**
 When you first deploy the executor, you should a single log line to indicate success:
@@ -147,6 +145,3 @@ Starting datahub executor worker
 ```
 This indicates that the remote executor has established a successful connection to your DataHub instance and is ready to execute ingestion & monitors.
 If you DO NOT see this log line, but instead see something else, please contact your Acryl representative for support.
-
-## Release Notes
-This is where release notes for the Acryl Remote Executor Container will live.
diff --git a/docs/ui-ingestion.md b/docs/ui-ingestion.md
@@ -338,7 +338,7 @@ for the `datahub-actions` container and running `docker logs <container-id>`.
 There are valid cases for ingesting metadata without the UI-based ingestion scheduler. For example,
 
 - You have written a custom ingestion Source
-- Your data sources are not reachable on the network where DataHub is deployed. Managed DataHub users can use a [remote executor](managed-datahub/operator-guide/setting-up-remote-ingestion-executor-on-aws.md) for remote UI-based ingestion.
+- Your data sources are not reachable on the network where DataHub is deployed. Managed DataHub users can use a [remote executor](managed-datahub/operator-guide/setting-up-remote-ingestion-executor.md) for remote UI-based ingestion.
 - Your ingestion source requires context from a local filesystem (e.g. input files)
 - You want to distribute metadata ingestion among multiple producers / environments
 

diff --git a/metadata-ingestion-modules/airflow-plugin/src/datahub_airflow_plugin/_extractors.py b/metadata-ingestion-modules/airflow-plugin/src/datahub_airflow_plugin/_extractors.py
@@ -262,7 +262,7 @@ def extract(self) -> Optional[TaskMetadata]:
         )
 
         operator: "BigQueryInsertJobOperator" = self.operator
-        sql = operator.configuration.get("query")
+        sql = operator.configuration.get("query", {}).get("query")
         if not sql:
             self.log.warning("No query found in BigQueryInsertJobOperator")
             return None

diff --git a/metadata-ingestion-modules/airflow-plugin/src/datahub_airflow_plugin/datahub_listener.py b/metadata-ingestion-modules/airflow-plugin/src/datahub_airflow_plugin/datahub_listener.py
@@ -247,10 +247,11 @@ def _extract_lineage(
                 SQL_PARSING_RESULT_KEY, None
             )
         if sql_parsing_result:
-            if sql_parsing_result.debug_info.error:
-                datajob.properties["datahub_sql_parser_error"] = str(
-                    sql_parsing_result.debug_info.error
-                )
+            if error := sql_parsing_result.debug_info.error:
+                logger.info(f"SQL parsing error: {error}", exc_info=error)
+                datajob.properties[
+                    "datahub_sql_parser_error"
+                ] = f"{type(error).__name__}: {error}"
             if not sql_parsing_result.debug_info.table_error:
                 input_urns.extend(sql_parsing_result.in_tables)
                 output_urns.extend(sql_parsing_result.out_tables)