Note: This is an article published in Linkedin published on 2019.

After one week, we are more than happy with the replacement of 23 old servers to 4 new.

The new infrastructure supports 170% more traffic than before using only 3 servers (the last server is connected, but working with only 5% of capacity and in standby mode for now).

In the last 2 months, we had the opportunity and time to:

  • Optimize the hardware, kernel, and system’s configurations on the new servers
  • Make different tests of stress, high-availability, disaster & recovery.
  • Discover and work with the development team to fix two bugs.

The sum of these 3 points, resulted in this spectacular performance.

Cluster Graphic Peformance. Cluster Graphic Peformance.

The graphic show very clear when a new server starts to work and yes, is “600 Gigabit/s” not “600 Megabit/s”. :-D

All this points mentioned before, show how important is to have the time and conditions to work well, before to put anything into production.