chore(updates.jenkins.io): activate HA for mirrorbits and httpd

[lemeurherve]

This PR set the replica count to 2 for the mirrorbits and the httpd subcharts of updates.jenkins.io to activate the High Availability of this service.

As the rsync server won't be requested as much as mirrorbits and Apache, I've kept its replica count to 1.

Ref:

• [INFRA-3100] Migrate updates.jenkins.io to another Cloud helpdesk#2649

[dduportal]

LGTM!

2 nitpicks (NOT blocking!)

1. I would change the title:

• This change is anew feature as it changes something for end users: requests are spread across replicas and the availability is increased
• The "component" should be publick8s or eventually publick8s/updates.jenkins.io to help the reviewer (or us in 3 months) to immediately locate which cluster was impacted. This one is a nit because we do not have a formal convention.

2. The helmfile diff only shows a Pod Disruption Budget for mirrorbits but not for the httpd component: worth checking once merged and deployed successfully (maybe it already exists, or maybe the httpd chart does not define a PDB when replicating). Might be worth checking along with get.jenkins.io

[dduportal]

LGTM!

2 nitpicks (NOT blocking!)

1. I would change the title:


* This change is  anew feature as it changes something for end users: requests are spread across replicas and the availability is increased

* The "component" should be `publick8s` or eventually `publick8s/updates.jenkins.io` to help the reviewer (or us in 3 months) to immediately locate which cluster was impacted. This one is a nit because we do not have a formal convention.


2. The helmfile diff only shows a Pod Disruption Budget for `mirrorbits` but not for the `httpd` component: worth checking once merged and deployed successfully (maybe it already exists, or maybe the httpd chart does not define a PDB when replicating). Might be worth checking along with get.jenkins.io

Issue: jenkins-infra/helm-charts#1109