Tolerex
Fault-tolerant distributed storage with leader/follower model and mTLS transport.
Architecture
Operational Metrics
| Metric | Observed Value | Target |
|---|---|---|
| Failover detection time | 2.1 s | < 3.0 s |
| Replication lag (median) | 130 ms | < 250 ms |
| Read availability (3-node) | 99.92% | > 99.9% |
Key Decisions
- Heartbeat and quorum checks separated from client request handler.
- mTLS for all node-to-node traffic.
- Persistent log checkpoints for predictable recovery.