Create new lagoon type "worker" for long operating processes

Currently, Lagoon projects that have long running (non-web) processes use the cli Lagoon type to run these processes in. This is not optimal, as the cli type should be reserved for short-running, instant spin-up type uses.

We are proposing to add an additional type worker and corresponding worker-persistent type to Lagoon. The main aim of these types is to accommodate projects that have long-running, persistent, and non-web-facing processes. Where a project contains a worker type and needs cronjobs - these cronjobs would be run inside the worker pod instead of a separate cron pod.

The main benefit of this is that cli pods can be idled when they are not in use, reducing load on a cluster, and can be treated appropriately during scaling or node-shuffling events. This will also mean that projects can operate their web-facing containers as lean as possible, offloading backend processing or interactive tasks to an additional pod, with specialist tooling.

[shreddedbacon]

Would be cool to be able to also have this be extendable with a custom port

services:
  worker:
    labels:
      lagoon.port.enabled: true
      lagoon.port.number: 8080
      lagoon.probe.liveness: 30
      lagoon.probe.readiness: 30

And these would be injected into the template with conditions for the port and direct for the probes. This would give some flexibility to the worker type then to be a little bit more generic.

For the probe overrides, it would probably be good to have some sane maximums.

This could also then possibly be used for #2528

[tobybellwood]

mhh how about we make it completely customizable and allow developers to inject helm variables via Lagoon?

Could this be applied to all helmcharts and not just this specific one?

The purpose of this chart is to define an always-running non-ingress type, instead of abusing the cli type - making this chart too flexible would leave it open to misuse as well!