S

System Resources Tracker for Cloud Instances

Cloud Providers... Assemble!

S

System Resources Tracker for Cloud Instances

Cloud Providers... Assemble!

The problem System Resources Tracker for Cloud Instances solves

Cloud is a concept that at first glance does not seem beginner-friendly. There are so many cloud providers in the market, each offering a plethora of services to their users. The problems beginners and system admins face include -

  • The dashboards provided by the cloud providers are complicated and take some time getting used to. The immediately look daunting to new users. Moreover, the dashboards are not uniform across providers as each provides different services in a different manner which makes it difficult to monitor those services

  • The abundance of data itself does not mean that meaningful insights will be captured. A good representation of data is necessary

  • If a user uses multiple instances from different providers then to access those instances the user has to access the shells through the provider's portal which becomes cumbersome if you need to jump between instances

  • As we scale an application, it becomes difficult to monitor and access services on different cloud providers

  • Extracting data from clusters that span multiple providers is still a tough ask

Our platform provides solutions to all the above problems. Monitor system metrics for all cloud services from one place. View CPU or memory usage, running processes and even access the terminal with a dead-simple dashboard.

The features included are -

  • Resource tracking of compute instances
  • Real-time data visualization of resources
  • Interactive shell access to the instance
  • Alerts when resources are critical
  • Multi-provider support
  • Easy Configuration

Challenges we ran into

  • Deciding and extracting meaningful metrics for the resources

  • A utility that tracks system resources should itself consume very few resources. It was a difficult task to efficiently track the resources and we succeeded in doing the same. On average, the data collection process requires < 2% of CPU usage

  • While deploying to cloud instances, we faced an issue related to the firewall

Discussion