O

Overclocked

Increasing productivity, one server at a time

O

Overclocked

Increasing productivity, one server at a time

The problem Overclocked solves

Absence of an organised and creative metrics analyser for servers.

In linux, there are two commands which serve the purpose of task manager; these are top and htop. However the output of these commands is not organised, neither interactive nor efficient in informing about all the processes.

The problems related to these commands are:

  • System Overload: You may lose access to your server during system overload
  • Time to live: The time to live for processes is very short and thus are often not displayed in the command output
  • Speed of instance: These commands are highly dependent on the speed of instance
  • Unresponsiveness: These commands don't work when the server is unresponsive
  • Bottlenecks: Bottlenecks are not easily identifiable using these commands

We came up with the solution to analyse all these metrics in the form of graphs, charts and illustrations.

The people it would help:

  • Sysadmins: Easier to monitor all the graphs and charts for a long time, using the primitive approach of top/htop would gradually increase the chance of error with time.
  • Cloud Services: Cloud services like AWS, Digital Ocean can provide their users with a dashboard like ours so that its easier to keep track of everything running and the health of your instance/server.
  • MNCs: MNCs have a lot of instances and a solution model like ours would provide an easier way to identify problems in their server.
  • Developers/DevOps Engineers: It would make the life of developers easier by notifications about specific parameters.

Challenges we ran into

  • We were using an AWS EC2 t2-micro instance to get the data of our server however due to very low RAM and and very low CPU our instance crashed two times and we had to rebuild it again thrice which would not have been the case if we were using an average type of instance instead of a very low tier one
  • We used AWS EC2 instances and linux and thus the only code we have in the github repository are the bash scripts we used to export metrics
  • We could not get the domain from MLH due to some issue with GoDaddy, we wanted to use the domain overclocked.biz

Discussion