[Week 9] Add Scaling to Your Projects #57

Open

opened 2026-05-08 13:09:04 +02:00 by balthazar · 2 comments

balthazar commented

2026-05-08 13:09:04 +02:00

Owner

1) Add Scaling to Your Projects

Do either of the following:

Create a high-availability setup with a hot and standby server for your ITU-MiniTwit application. In such a setup you would have two replica of your application server where only one is ever active at a time, i.e., no load balancer in front of either of them.
The setup should resemble the one discussed in the lecture.

Or:

Create a Docker Swarm cluster for your ITU-MiniTwit applications in which various components run as services and can be scaled as needed. Note, very likely you do not want to scale you DB servers horizontally (to keep your data consistent) but only all the services 'in front' of it.

1.b) (Optional) Measure your downtime when migrating to the swarm

How much time is your system down? Stopwatch measurement is fine. Ideally zero downtime should be achievable if you don't have to move your database.
A brief description (half a page to one page) of how you ensured a low downtime during this transition should be part of your final report

## 1) Add Scaling to Your Projects Do _either_ of the following: 1. Create a high-availability setup with a _hot_ and _standby_ server for your _ITU-MiniTwit_ application. In such a setup you would have two replica of your application server where only one is ever active at a time, i.e., no load balancer in front of either of them. The setup should resemble [the one discussed in the lecture](https://www.digitalocean.com/community/tutorials/how-to-set-up-highly-available-web-servers-with-keepalived-and-floating-ips-on-ubuntu-14-04). Or: 2. Create a Docker Swarm cluster for your _ITU-MiniTwit_ applications in which various components run as services and can be scaled as needed. Note, very likely you do not want to scale you DB servers horizontally (to keep your data consistent) but only all the services 'in front' of it. ## 1.b) (Optional) Measure your downtime when migrating to the swarm - How much time is your system down? Stopwatch measurement is fine. Ideally zero downtime should be achievable if you don't have to move your database. - A brief description (half a page to one page) of how you ensured a low downtime during this transition should be part of your final report