EPIC: Maintenance pages

[mrnicegyu11]

To have a maintenance page even if the docker swarm is down or machines are down, a machine hosting the nginx-based maintenance page needs to be added to the load-balancing.

On aws:

• Add a dedicated ultra-small machine, or even better, a lambda, which only shows the maintenance page. This is added to the osparc.io and staging.osparc.io loadbalancer.

On dalco:

• Request a VM, or if impossible, a shitty old machine
• Provision with keepalived (as found in osparc-infra/ansible) to share the Floating IP
• Have it run only the maintenance page

On master:

• We dont need it for now

Tasks

Beta Give feedback

1. Fix broken keepalived FloatingIP maintenance page for dalco or tip #325
   during maintenance p:mid-prio t:enhancement +3
2. Bug: Only one maintenance page possible in dalco #214
   FAST p:mid-prio t:bug wontfix +4
3. Add fallback nginx to serve "something went wrong" page #218
   p:mid-prio t:enhancement +2
4. Test load-balancing maintenance page #409
   0 of 4
   during maintenance p:high-prio +2
5. Add FloatingIP maintenance page for dalco with VM from O+P
6. Audit ansible provisioning of maintenance VM machines

[YuryHrytsuk]

This is done for NIH / DALCO / TIP / Sim4life PROD