Skip to content

Latest commit

 

History

History
74 lines (59 loc) · 3.33 KB

README.md

File metadata and controls

74 lines (59 loc) · 3.33 KB
Raw

Multi-Hauser with Docker

If your FullStory account has multiple associated accounts, this recipe can be used as a template to start a docker container that exports the data for each "org" and uploads it to BigQuery. While this is intended to work with Google Cloud, it can easily be adapted to work with AWS or any databases supported in the future.

How it works

This recipe works by creating a docker container that uses supervisord to run multiple hauser processes against multiple FullStory accounts (i.e. "orgs"). When the docker container is started, it uses the provided key file to build a config.toml for each org from the hauser config template and saves it to a working directory for that org (e.g. /hauser/<orgId>/config.toml). A "program" is also added to the supervisor config file. Supervisor is then started and runs in the foreground until the container is terminated.

Hauser Configuration Template

The template in this directory requires that three variables are defined:

  • Bucket: Google Cloud Storage bucket ([gcs] section)
  • Project: Google Cloud project ID ([bigquery] section)
  • Dataset: BigQuery dataset that the tables will be created in ([bigquery] section)

The other hauser configuration variables can be customized as well, but the above are the minimum required.

Export Delay

In an attempt to offset the times at which each process is doing work, the generated hauser configs will have a 10 min offset in their ExportDelay value. So, if your key file has 3 lines, the ExportDelay value for each org will be 24 hours, 24 hours and 10 minutes, and 24 hours and 20 minutes, respectively.

Key File

The key file should be a CSV with the following format:

<org1>,<api_key_for_org1>
<org2>,<api_key_for_org2>
<org3>,<api_key_for_org3>
...

Important: The last line must have a new line at the end or it will be dropped.

Building and running the docker container

Assuming the current working directory is the directory of this README, the docker container can be built with the following command:

docker build -t multi-hauser . --build-arg HAUSER_VERSION=1.0.0

This builds a docker image that does not contain any sensitive information (such as API keys and GCP credentials), and thus is safely portable/shareable.

Running the container is a bit more involved and can change based on how the deployment is being managed. For instance, with Google Cloud, the container should use a service account so that hauser can make API calls to Google Cloud Storage and BigQuery. In Google Kubernetes Engine (GKE), the credentials can easily be stored in secrets and mounted to the docker container as specified in the deployment. If running on a VM, the credentials can just be copied directly to the the desired location.

The following docker run command is just an example for running the process manually and assumes that the key file and service account credentials are in the current directory.

docker run --rm -it \
  -v ${PWD}/keys.csv:/secrets/keys.csv \
  -v ${PWD}/account.json:/secrets/account.json \
  -e GOOGLE_APPLICATION_CREDENTIALS=/secrets/account.json \
  --name multi-hauser
  multi-hauser:latest /secrets/keys.csv