Skip to content

Latest commit

 

History

History
130 lines (90 loc) · 3.35 KB

README.md

File metadata and controls

130 lines (90 loc) · 3.35 KB
Raw

GCP Log Manager

GCP Log Manager is a tool designed to collect app logs from devices and provide analytics about daily user statistics. This program simplifies the process of managing logs using various Google Cloud Platform (GCP) services.

System Overview

  1. The GCP Log Manager collects app logs from devices and provides analytics about daily user statistics.
  2. The workflow involves event logs sent to the /api/postLog endpoint, which are published to Google Cloud Pub/Sub.
  3. A Dataflow job then populates a Google BigQuery table, and the /api/getAnalytics API call queries BigQuery to retrieve daily user statistics.

This solution streamlines log management, facilitates data analysis, and empowers users to make informed decisions based on log data.

Installation and Setup

.env Setup

Place create .env files on both server and scripts directory by following the .env.sample format with GCP account specific values.

GCP Service Account Credentials

Place your GCP service account credentials JSON file inside the project directory (./gcp-log-manager). Name the JSON file gcp-service-account.json.

Running the Application

Dockerized Version

To run the GCP Log Manager as a Docker container, follow these steps:

  1. Make the start script executable:
chmod +x ./start.sh
  1. Run the start script to build the Docker image, start the application container, and manage the Dataflow job:
./start.sh
  1. Prompt exit to close the application and remove the container/image

Non-Dockerized Version

To run the GCP Log Manager without Docker, follow these steps:

  1. Make the start script and related scripts executable:
chmod +x ./scripts/startDataflow.sh ./scripts/cancelDataflow.sh
  1. Install the required dependencies by running the following command:
npm install
  1. Start the application using Node.js:
cd ./server; node ./index.js
  1. You can stop the application with Command + C (CTRL + C) command.

API Endpoints

  1.  `POST localhost:8080/api/postLog`

This endpoint is used to send logs to the GCP Log Manager. The expected input format is as follows:

{
 "type": "$type",
 "session_id": "$session_id",
 "event_name": "$event_name",
 "event_time": $event_time,
 "page": "$page",
 "country": "$country",
 "region": "$region",
 "city": "$city",
 "user_id": "$user_id"
}

This endpoint returns a success message as a response.

  1. GET localhost:8080/api/getAnalytics

This endpoint provides aggregated daily user statistics. It returns data in a specific format as follows:

{
   "total_users": 916,
   "daily_stats": [
         {
            "date": "2022-01-16",
            "average_session_duration": 304,
            "active_user_count": 267,
            "new_user_count": 10
        },
        ...
      ]
}

Using Helm for Deployment

  1. Build and push the docker image of this project to an image repository.
  2. Change the values of the "image" variables on ./helm/deployement.yaml and ./helm/templates/values.yaml
  3. Adjust deployment and service settings.
  4. Deploy the service with:
helm install gcp-log-manager .

Dependencies

The GCP Log Manager uses the following packages:

  • @google-cloud/bigquery: "^7.2.0"
  • @google-cloud/dataflow: "^3.0.0"
  • @google-cloud/pubsub: "^4.0.1"
  • body-parser: "^1.20.2"
  • dotenv: "^16.3.1"
  • express: "^4.18.2"
  • shelljs: "^0.8.5"