fix(config): reduced preprocessing replicas from 2 to 1

[tombch]

Discussed with Anya - we said it looks like the replicas for Loculus preprocessing should be one instead of two.

PR Checklist

• All necessary documentation has been adapted.
• The implemented feature is covered by appropriate, automated tests.
• Any manual testing that has been done is documented (i.e. what exactly was tested?)

🚀 Preview: Add preview label to enable

[claude]

Claude finished @tombch's task in 40s —— View job

---

Code Review

• Gather context and read changed files
• Review the change

---

This is a minimal, focused change — reducing the default preprocessing replicas from 2 to 1 in kubernetes/loculus/values.yaml.

Summary: The change is straightforward and correct in isolation. A few observations worth noting:

• No test coverage needed — this is a configuration default, not application logic.
• No documentation impact — the change is self-explanatory.
• Missing PR context — the checklist items are unchecked and there's no linked issue. It would be helpful to briefly explain why this was reduced (e.g. resource constraints, over-provisioning, observed idle replicas) so future readers understand the intent.
• Potential concern — reducing from 2 to 1 removes redundancy for the preprocessing service. If this is a production default, consider whether a single replica is acceptable for availability. If this is only for local/dev defaults or to reduce resource usage in shared environments, that's reasonable — but worth calling out explicitly.

The helm lint validation mentioned in kubernetes/CLAUDE.md should be run to confirm the values file is still valid after the change.