Skip to content

Latest commit

 

History

History
69 lines (50 loc) · 2.53 KB

README.md

File metadata and controls

69 lines (50 loc) · 2.53 KB

Slack to Snowflake Fivetran Pipeline Generator

Overview

System architecture diagram

Required Environment Prerequisites

  • Must have an AWS account user and associated access key credentials with permissions to:
    • Create and modify lambda functions
    • Create and modify IAM roles and policies
    • Create S3 buckets
    • Write to a pre-defined terraform state S3 bucket (if remote terraform state is desired)
  • Must have a Snowflake account with a Snowflake user (and password) with permissions (see sample setup script):
    • USAGE on a warehouse for executing queries
    • Granted a role that has permissions to create users/warehouses/databases
  • Must have a Fivetran account and an associated API key/secret pair
  • Must have a Slack account and an associated token with permissions to the resources that will be pulled by the lambda function

Sample Snowflake Setup Script

CREATE USER TERRAFORM_USER PASSWORD='<password>';
CREATE ROLE TERRAFORM_ROLE;
CREATE WAREHOUSE TERRAFORM_WH;
GRANT ROLE ACCOUNTADMIN TO ROLE TERRAFORM_ROLE;
GRANT ROLE TERRAFORM_ROLE TO USER TERRAFORM_USER;
GRANT ALL PRIVILEGES ON WAREHOUSE TERRAFORM_WH TO ROLE TERRAFORM_ROLE;

Usage

This is currently meant to be executed locally using the run.local.sh. To adapt this to running in a CICD environment - variables, secrets, and terraform state would need to be managed by the CICD tool instead of using the .env file.

Lambda Code

Update the aws/lambda/function.py python script with the code that will be used to query slack data and pass it back to Fivetran

Infrastructure Naming

Update the values in main.tf if desired to change the names of the infrastructure that will be created

Environment Configuration

Create a file in the root directory of this repository named .env and paste the following into it:

# These are all dummy values - replace with your own

# Optional depending on whether you want to use remote state management or not
TF_STATE_BUCKET="bucket-name"

AWS_ACCESS_KEY_ID="key-id"
AWS_SECRET_ACCESS_KEY="access-key"
AWS_REGION="us-east-1"

FIVETRAN_APIKEY="api-key"
FIVETRAN_APISECRET="secret"

SNOWFLAKE_USER="TERRAFORM_USER"
SNOWFLAKE_PASSWORD="password"
SNOWFLAKE_ACCOUNT="xxxxxx"
SNOWFLAKE_REGION="us-east-2.aws"
SNOWFLAKE_WAREHOUSE="TERRAFORM_WH"
SNOWFLAKE_ROLE="ACCOUNTADMIN"

SLACK_TOKEN="token"

Execution and Deployment

Execute the run.local.sh script and follow the prompts to deploy the infrastructure