Increase memory limit for MCP server containers to 2Gi

[aponcedeleonch]

Based on user feedback in Discord

Updated memory limits across all Helm configurations and documentation to address out-of-memory issues observed in production deployments.

Changes:

• Increased memory limit from 1Gi to 2Gi in default values.yaml
• Updated development and single-namespace example configurations
• Modified documentation and example manifests to reflect new limits
• Memory requests remain unchanged at 256Mi (development) and 128Mi (examples)

[claude]

Code Review

Summary

Good straightforward change addressing production OOM issues. The increase from 1Gi to 2Gi is well-documented and justified by user feedback.

Issues Found

Missing Updates (High Priority)
Two config files still have 1Gi memory limits:

• helm/mcp-optimizer/examples/values-with-group-filtering.yaml:15 - should be 2Gi for consistency
• local-values.yaml:53 - should be 2Gi (or document why it's intentionally different)

Inconsistency (Medium Priority)
values-development.yaml and values-single-namespace.yaml jumped from 512Mi to 2Gi (4x increase), while values.yaml went from 1Gi to 2Gi (2x increase). Consider if development configs need the full 2Gi or if 1Gi would suffice.

Potential Concern
values-with-persistence.yaml:18 has memory request at 512Mi (vs 256Mi/128Mi elsewhere). Verify this is intentional.

Recommendations

• Update the two missing files to 2Gi
• Document rationale if different limits are intentional for specific environments
• Consider whether development environments need full 2Gi or can use 1Gi

Best Practices ✅

• All documentation updated
• Changes are minimal and focused
• Clear commit message referencing user feedback