Investigate the cause of lengthy tests and video in E2E

[cnotv]

Description

Some tests (e.g. c1667678-cb0f-491e-b9cd-51f996550d03, 5373bf26-386f-4607-9af5-80789a08369d) seem to invest specifically an extensive amount of time without any tracked waiting from the network.
The problem seems to do not happen locally, as typical scenario opposed to reproducing E2E tests in CI.

This issue cause generation of lengthy videos which increase costs on our service.

Automation consist in reproducing the same test in a shorter time.

[cnotv]

I noticed that we are running Node 14 instead of 16, which makes everything slower.
@richard-cox could you please confirm it to me?

This is the timing in CI on push on my fork and it seems better:
https://github.com/cnotv/dashboard/actions/runs/5078845852/jobs/9123864155
As well as the report in our Cypress dashboard: bae8360287c5b153b39bb9fee92a11c8

[cnotv]

Never mind, just a coincidence.

[gaktive]

Perhaps the videos can be detached from the droplet and placed into a storage bucket. They are useful for detecting regressions or changes in behaviour.