Skip to content

service that executes queries on BigQuery datasets generated by Helix-Logging

License

Notifications You must be signed in to change notification settings

adobe/helix-run-query

Helix Run Query

A service that runs premade queries on datasets created by Helix-Logging

Status

codecov CircleCI GitHub license GitHub issues LGTM Code Quality Grade: JavaScript semantic-release

Usage

Assuming you are using curl as your http client:

curl https://helix-pages.anywhere.run/helix-services/run-query@v3/rum-dashboard
{
    "results": {
      "data": [
        {
            "req_url": "https://helix-secret.fake",
            "resp_http_Content_Type": "text/html; charset=UTF-8",
            "status_code": "404",
            "visits": 1
        }
      ],
    }
}

The data is returned in Franklin's Spreadsheet Object Notation (SSHON) format. For more information, see Franklin's documentation. There are two sheets, one called meta with information about the query and the other called results with the actual data.

Authentication

The service is protected by shared secrets that apply to a domain and all subdomains. These secrets are called domain keys.

You can pass the domain key in two ways:

  • as a query parameter ?domainkey=…
  • as an Authorization: Bearer … header

Domain keys can be read-only or read-write. Read-write domain keys allow you to issue new domain keys and revoke exiting keys. They should never be passed to the service in a GET request.

It is best practice to use read-write domain keys only to issue new read-only keys and work with those.

Query Parameters

Each query can define a set of query parameters, but the following parameters are supported almost universally:

  • limit - the maximum number of rows to return
  • interval – the number of days of data to return
  • offset – the number of rows or days to skip
  • startdate – the start date of the data to return (this will only take effect if interval is not set)
  • enddate – the end date of the data to return (this will only take effect if interval is not set)
  • timezone – the timezone to use for the dates (defaults to UTC)
  • url – the URL to filter the data by. By default, the URL is taken as a partial filter, so url=https://www.adobe.com/ will include https://www.adobe.com/, https://www.adobe.com/?campaign=exapmle, and https://www.adobe.com/en/. If you want to filter by an exact URL, use url=https://www.adobe.com/$. If you want to filter by an exact URL, but allow URL parameters, use url=https://www.adobe.com/?
  • checkpoint – filter results by a particular checkpoint. The list of common checkpoints can be found here.
  • device – filter results by a particular class. Possible values are all (default), desktop, mobile, and bot.
  • source – filter results by a particular source parameter value

Deprecation Notice

helix-run-query@v2 is no longer available. Please use helix-run-query@v3 instead.

Changes from v2 to v3

  • the output format has changed from a custom format to Franklin's Spreadsheet Object Notation (SSHON)
  • the domain parameter has been removed and replaced with url
  • the url parameter now supports partial matching and exact matching

For a full list of changes, please see the CHANGELOG.

Development

dev.sh script

Running queries in development using the bq tool can be cumbersome, as all parameters in the query are required. The dev.sh script can be used to run queries in development mode and will fill out important parameters with defaut values.

Start setting it up by exporting your domain key to DOMAINKEY, then run dev.sh <query>.

$ export DOMAINKEY=…
$ sh dev.sh rum-dashboard url www.adobe.com

You can pass additional parameters to the query as seen above. The second (optional) parameter allows you to specify the output format, with values pretty, csv, json being supported.

Required Environment Variables

This service depends on three external services to operate:

  • Fastly
  • Google Cloud Platform

It is configured using a number of environment variables that are required for testing (tests that miss required variables will be skipped) and deployment (deployment will fail or be non-functional). These variables are required and this is how to set them up:

GOOGLE_CLIENT_EMAIL

This is the email address associated with a Google Cloud Platform Service account. It looks like <name>@<project>.iam.gserviceaccount.com. You can create a proper service account following the instructions in the Google Cloud Platform documentation or this step-by-step guide:

  1. Log in to Google Cloud Platform Console
  2. Select menu → "IAM & admin" → "Service accounts" → "Create service account"
  3. Create the service account
  4. Add the following roles to the service account:
    • BigQuery Admin
    • Service Account Admin
    • Service Account Key Admin
    • Service Account Key Admin
  5. Create a private key in JSON format for the service account and download the key file

Note: The private key file and the value of the GOOGLE_CLIENT_EMAIL environment variable should be considered private and should never be checked in to source control.

The downloaded file will look something like this:

{
  "type": "service_account",
  "project_id": "project-12345678",
  "private_key_id": "111122223333aaaabbbbccccdddd123412345",
  "private_key": "-----BEGIN PRIVATE KEY-----\n\n-----END PRIVATE KEY-----\n",
  "client_email": "[email protected]",
  "client_id": "111122223333444456789",
  "auth_uri": "https://accounts.google.com/o/oauth2/auth",
  "token_uri": "https://oauth2.googleapis.com/token",
  "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
  "client_x509_cert_url": "https://www.googleapis.com/robot/v1/metadata/x509/example-account%40project-12345678.iam.gserviceaccount.com"
}

Copy the value of the client_email field (e.g. [email protected]) and save it in the GOOGLE_CLIENT_EMAIL environment variable.

GOOGLE_PRIVATE_KEY

This is the private key associated with the Google Cloud Platform Service account created above. In order to retrieve the correct value, see Creating and Managing Service Account Keys in the Google Cloud Platform documentation or continue the step-by-step guide from above:

  1. Make sure you've followed all steps to get the value of GOOGLE_CLIENT_EMAIL
  2. Copy the value of the private_key property in the JSON file you've downloaded

Note: The private key and the value of the GOOGLE_PRIVATE_KEY environment variable should be considered private and should never be checked in to source control.

The private key is a multi-line value.

Note: Private keys created using an API typically have a short expiration time and need to be rotated in regular intervals. Even for private keys that have been created manually, regular rotation is a best practice.

GOOGLE_PROJECT_ID

This is the Google Cloud Platform project ID. It looks like project-12345678 and you will find it in lots of places in the Google Cloud Platform Console UI. In addition, you can just take the value of the project_id property in your downloaded key JSON file.

HLX_FASTLY_NAMESPACE

This property is only required for testing and development. It is the service config ID that you can retrieve from Fastly.

For testing, it is a good idea to use a separate, non-production service config, as the tests not only perform frequent updates, but they also rotate the private keys of the created Google Cloud Platform service accounts. As the tests don't activate the service config, this will lead to an invalid logging configuration in a short time.

HLX_FASTLY_AUTH

This property is only required for testing and development. It is an API token for the Fastly API. Follow the instructions in the Fastly documentation to create a token.

The token needs to have global, i.e. write access to your service config.

Note: The API token and the value of the HLX_FASTLY_AUTH environment variable should be considered private and should never be checked in to source control. For more, see the API documentation.

Development

You need node>=8.0.0 and npm>=5.4.0. Follow the typical npm install, npm test workflow.

Contributions are highly welcome.

Queries

Query files live in the src/queries directory, although they can also live in subdirectories of src/queries. They are static resources; that are loaded into run-query and then sent to BigQuery for actual execution. In order to be bundled into the lambda function, they have to be listed in package.json It is up to the developer to ensure their query is correct; this can be done by using the BigQuery console.

The build process includes a sqlfluff check. The recommendation is to install sqlfluff locally so that you can fix any syntax issues prior to commit. See https://docs.sqlfluff.com/en/stable/gettingstarted.html.

Once a query file is complete and correct, you may add it as a static resource; so that it won't be excluded during OpenWhisk deployment. In the root of the repository; find the package.json and add your query file (file with .sql extension) under static:

"wsk": {
"name": "helix-services/run-query@${version}",
"static": [
     "src/queries/next-resource.sql"
  ]
},

Now, query file can be executed as an action; triggered by a request as such:

curl -X POST -H "Content-Type: application/json" https://adobeioruntime.net/api/v1/w/helix-services/run-query@v1/next-resource -d '{"service":"secretService", "token":"secretToken", "queryParam1":"value"}'

Parameterized Queries

Helix Run Query provides query developers the ability to specify parameters anywhere in their queries. Using ^param anywhere in the query; and providing a corresponding {param: 'value'} in the request; enables you to effectively parameterize just about any part of the query. It is not advised to parameterize anything beyond table name or limit. Anything else; can make your query susceptible to SQL injection.

Deploying Helix Run Query

Deploying Helix Run Query requires the wsk command line client, authenticated to a namespace of your choice. For Project Helix, we use the helix namespace.

All commits to main that pass the testing will be deployed automatically. All commits to branches that will pass the testing will get commited as /helix-services/run-query@ci<num> and tagged with the CI build number. The CI build number can be found next to the branch-deploy job for the pipeline execution related to a commit at https://app.circleci.com/pipelines/github/adobe/helix-run-query.

For a query to be available at helix-services/run-query@v3 (semantic versioning), a commit in the pull request must follow the message convention standards defined at https://www.conventionalcommits.org/en/v1.0.0/.