Increase default worker-high resourcing

[derek-etherton-opslevel]

Bump default opslevel.workerHigh.resources from small (200m CPU / 500Mi memory requests) to medium (1 CPU / 2Gi memory requests) so high-priority Sidekiq workers have enough CPU and memory to avoid restarts and degraded deployments on typical self-hosted installs.

Problem

We are seeing degraded performance of one customer's worker-high deployment.

Solution

We already use resourcesMedium for our worker-low deployment, there's no reason we should be allocating less to our high priority workloads. Bumps resourcing from small (200m CPU / 500Mi memory requests) to medium (1 CPU / 2Gi memory requests).

I had AI help outline our current resource requests:

Component	Pods	CPU (req)	Memory (req)
OpsLevel web	3	3	6 Gi
worker-high	2	2 (was 0.4)	4 Gi (was 1 Gi)
worker-low	2	2	4 Gi
scheduler	1	1	2 Gi
worker-faktory	1	1	2 Gi
worker-search	1	1	2 Gi
Opssight web	1	1	2 Gi
Opssight worker	1	0.2	0.5 Gi
MySQL	1	1	2 Gi
Postgres	1	1	2 Gi
Redis	1	1	2 Gi
Elasticsearch	1	2	2 Gi
MinIO	4	4	8 Gi
Faktory	1	1	2 Gi
Total	21	~21.2 (was ~19.6)	~40.5 Gi (was ~37.5 Gi)

Comparing this back to our documented Resource Requirements:

The Kubernetes cluster must meet the following minimum specifications:

    CPU: 16 cores
    Memory: 20 GB RAM
    Pods: Capacity for at least 20 pods
    Disk Storage: Minimum of 120 GiB (may be lower if the production database is hosted elsewhere)

It seems like our recommended minimums have been well below the current requirements. Am I missing something here, or do the docs need updated?

[jasonopslevel]

docs definitely need updating!