airbyte-serverless 0.23

Airbyte made easy (no UI, no database, no cluster)

Airbyte made simple

---

🔍️ What is AirbyteServerless?

AirbyteServerless is a simple tool to manage Airbyte connectors, run them locally or deploy them in serverless mode.

💡 Why AirbyteServerless?

Airbyte is a must-have in your data-stack with its catalog of open-source connectors to move your data from any source to your data-warehouse.

To manage these connectors, Airbyte offers Airbyte-Open-Source-Platform which includes a server, workers, database, UI, orchestrator, connectors, secret manager, logs manager, etc.

AirbyteServerless aims at offering a lightweight alternative to Airbyte-Open-Source-Platform to simplify connectors management.

📝 Comparing Airbyte-Open-Source-Platform & AirbyteServerless

Airbyte-Open-Source-Platform	AirbyteServerless
Has a UI	Has NO UI Connections configurations are managed by documented yaml files
Has a database	Has NO database - Configurations files are versioned in git - The destination stores the state (the checkpoint of where sync stops) and logs which can then be visualized with your preferred BI tool
Has a transform layer Airbyte loads your data in a raw format but then enables you to perform basic transform such as replace, upsert, schema normalization	Has NO transform layer - Data is appended in your destination in raw format. - airbyte_serverless is dedicated to do one thing and do it well: Extract-Load.
NOT Serverless - Can be deployed on a VM or Kubernetes Cluster. - The platform is made of tens of dependent containers that you CANNOT deploy with serverless	Serverless - An Airbyte source docker image is upgraded with a destination connector - The upgraded docker image can then be deployed as an isolated Cloud Run Job (or Cloud Run Service) - Cloud Run is natively monitored with metrics, dashboards, logs, error reporting, alerting, etc - It can be scheduled or triggered by events
Is scalable with conditions Scalable if deployed on autoscaled Kubernetes Cluster and if you are skilled enough. 👉 Check that you are skilled enough with Kubernetes by watching this video 😁.	Is scalable Each connector is deployed independently of each other. You can have as many as you want.

💥 Getting Started with abs CLI

abs is the CLI (command-line-interface) of AirbyteServerless which facilitates connectors management.

Install abs 🛠️

pip install airbyte-serverless

Create your first Connection 👨‍💻

abs create my_first_connection --source="airbyte/source-faker:0.1.4" --destination="bigquery" --remote-runner "cloud_run_job"

1. Docker is required. Make sure you have it installed.
2. source param can be any Public Docker Airbyte Source (here is the list). We recomend that you use faker source to get started.
3. destination param must be one of the following:
   • print (default value if not set)
   • bigquery
   • contributions are welcome to offer more destinations 🤗
4. remote-runner param must be cloud_run_job. More integrations will come in the future. This remote-runner is only used if you want to run the connection on a remote runner and schedule it.
5. The command will create a configuration file ./connections/my_first_connection.yaml with initialized configuration.
6. Update this configuration file to suit your needs.

Run it! ⚡

abs run my_first_connection

1. This will launch an Extract-Load Job from the source to the destination.
2. The run commmand will only work if you have correctly edited ./connections/my_first_connection.yaml configuration file.
3. If you chose bigquery destination, you must:
   • have gcloud installed on your machine with default credentials initialized with the command gcloud auth application-default login.
   • have correctly edited the destination section of ./connections/my_first_connection.yaml configuration file. You must have dataEditor permission on the chosen BigQuery dataset.
4. Data is always appended at destination (not replaced nor upserted). It will be in raw format.
5. If the connector supports incremental extract (extract only new or recently modified data) then this mode is chosen.

Select only some streams 🧛🏼

You may not want to copy all the data that the source can get. To see all available streams run:

abs list-available-streams my_first_connection

If you want to configure your connection with only some of these streams, run:

abs set-streams my_first_connection "stream1,stream2"

Next run executions will extract selected streams only.

Handle Secrets 🔒

For security reasons, you do NOT want to store secrets such as api tokens in your yaml files. Instead, add your secrets in Google Secret Manager by following this documentation. Then you can add the secret resource name in the yaml file such as below:

source:
  docker_image: "..."
  config:
    api_token: GCP_SECRET({SECRET_RESOURCE_NAME})

Replace {SECRET_RESOURCE_NAME} by your secret resource name which must have the format: projects/{PROJECT_ID}/secrets/{SECRET_ID}/versions/{SECRET_VERSION}. To get this path:

1. Go to the Secret Manager page in the Google Cloud console.
2. Go to the Secret Manager page
3. On the Secret Manager page, click on the Name of a secret.
4. On the Secret details page, in the Versions table, locate a secret version to access.
5. In the Actions column, click on the three dots.
6. Click on 'Copy Resource Name' from the menu.

Run from the Remote Runner 🚀

abs remote-run my_first_connection

2. The remote-run commmand will only work if you have correctly edited ./connections/my_first_connection.yaml configuration file including the remote_runner part.

1. This command will launch an Extract-Load Job like the abs run command. The main difference is that the command will be run on a remote deployed container (we use Cloud Run Job as the only container runner for now).
2. If you chose bigquery destination, the service account you put in service_account field of remote_runner section of the yaml must be bigquery.dataEditor on the target dataset and have permission to create some BigQuery jobs in the project.
3. If your yaml config contains some Google Secrets, the service account you put in service_account field of remote_runner section of the yaml must has read access to the secrets.

Schedule the run from the Remote Runner ⏱️

abs schedule-remote-run my_first_connection "0 * * * *"

⚠️ THIS IS NOT IMPLEMENTED YET

Get help 📙

$ abs --help
Usage: abs [OPTIONS] COMMAND [ARGS]...

Options:
  --help  Show this message and exit.

Commands:
  create                  Create CONNECTION
  list                    List created connections
  list-available-streams  List available streams of CONNECTION
  remote-run              Run CONNECTION Extract-Load Job from remote runner
  run                     Run CONNECTION Extract-Load Job
  run-env-vars            Run Extract-Load Job configured by environment...
  set-streams             Set STREAMS to retrieve for CONNECTION (STREAMS...

Keep in touch 🧑‍💻

Join our Slack for any question, to get help for getting started, to speak about a bug, to suggest improvements, or simply if you want to have a chat 🙂.

👋 Contribute

Any contribution is more than welcome 🤗!

• Add a ⭐ on the repo to show your support
• Join our Slack and talk with us
• Raise an issue to raise a bug or suggest improvements
• Open a PR! Below are some suggestions of work to be done:
  • implements a scheduler
  • implement the get_logs method of BigQueryDestination
  • enable updating cloud run job instead of deleting/creating when it already exists
  • add a new destination connector (Cloud Storage?)
  • add more remote runners such compute instances.
  • implements vpc access
  • implement optional post-processing (replace, upsert data at destination instead of append?)

🏆 Credits

• Big kudos to Airbyte for all the hard work on connectors!
• The generation of the sample connector configuration in yaml is heavily inspired from the code of octavia CLI developed by airbyte.