docs(readme): document clustering and add replication/HA roadmap - #185

summary

the readme's clustering section existed but didn't explain what the cluster actually does or show how to start one manually. this PR fills in the gaps and adds the replication/HA roadmap now that the cluster module is architecturally complete.

changes:

• new one-line summary of how clustering works (16384 hash slots, SWIM gossip, Raft topology, MOVED/ASK)
• manual 3-node quickstart alongside the make cluster shortcut, showing --cluster-enabled, --cluster-bootstrap, CLUSTER MEET, and slot assignment
• full cluster command reference table (15 commands)
• note that REPLICATE and FAILOVER are coming next — with a pointer to the status table
• four new cluster config flags in the configuration table (--cluster-enabled, --cluster-bootstrap, --cluster-port-offset, --cluster-node-timeout)
• architecture section now describes cluster mode alongside sharded and concurrent modes
• status table updated: phase 5 stays 🚧 (CLI and observability done, clients not started), phase 6 added as 🔜 with a brief description of what remains (replication stream, role propagation, failover)
• stats updated: 1,049 tests, ~22k LOC (~41k including tests/comments), 107 commands

what was tested

• cargo build --workspace — clean build
• cargo clippy --workspace -- -D warnings — zero warnings
• cargo fmt --check — clean (one unrelated formatting fix in integration tests included)
• read through the full clustering section for accuracy and flow

design considerations

the manual quickstart uses $(seq 0 5460 | tr '\n' ' ') for slot assignment. this is realistic for documentation — in practice make cluster handles it. the replication roadmap is kept brief in the README (one paragraph + status row) since the detailed design lives in the implementation plan.