Solution for projects that are time sensitive and have high traffic peaks on specific timeslots

Various projects had high traffic peaks in very specific timeslots compared with their regular traffic (100 times increase in the timespan of an hour, for example). I looked into what is the best approach to continue providing the service to the end users while being economically feasible.

Upgrading the server out right would be the easy choice, however this can get expensive very fast if the peak traffic only happens for a short amount of time.

As such, AWS was configured to launch X amount of machines whenever the traffic increased (which can be measured directly via the CPU usage of each instance).

All this was handled in the code (Infrastructure as Code) with the help of Terraform. The machines were all created using the same template with the help of Ansible.

When the peak traffic was over, the extra unneeded instances would be teared down. The end users would notice no difference other than the applications running smoothly no matter the traffic.

If you’d like a similar set up, send me a message and let’s see how I can help you!