Update 10k example chart docs to latest recommended values

What does this MR do?

MR updates the example 10k values chart to match findings from performance testing and what we now use in GET.

While adding in Hybrid Reference Architecture support into GET it was found that the example 10k chart values didn't give comparable Webservice performance to the standard 10k Reference Architecture. After a significant amount of testing this was due to a smaller amount of total Puma workers (56 versus 78 in the normal architecture) along with finding that Puma performed better in 4 worker pods instead of 2.

In addition to this the following tweaks are made:

• Recommend 20 pods with 4 workers each instead of 28 pods with 2 workers for a total worker count increase along with improved performance
• Use the default Puma workerMaxMemory of 1024 instead of overriding. In Omnibus 1024 is the current recommended default and this should be followed. If adjustment is required the defaults in both Charts and Omnibus should be adjusted instead.
  • With the smaller memory requirements adjust the Requests and Totals count to fit more pods with limits in the recommended Node Pool size. Recommend 1.125GB per worker now.
  • Thread count is also recommended to be no higher than 4 as the benefits after this value were found to be minimal after extensive testing.
  • With the above the whole section can be removed as we're now following the defaults.
• Remove the Workhorse extraArgs section. Same as Puma memory this should be adjusted in the product itself rather than being overridden by the users in all cases. A separate issue is opened to explore this.
• Add in more sections for Secrets and Object Storage buckets to be clearer

Related issues

Relates gitlab-org/quality/reference-architectures#30 (closed) https://gitlab.com/gitlab-org/quality/gitlab-environment-toolkit/-/issues/79

Closes gitlab-org/quality/reference-architectures#41 (closed)