#RF-BLOG-04: Keeping Decentralized Identity Servers Always-On — Async Tasks, Failover, and Circuit Breakers in Rust

## #RF-BLOG-04 — Keeping Decentralized Identity Servers Always-On

**Type:** Child Story / Blog Post
**Parent:** Epic #515 (#RF-BLOG-01)
**Status:** Ready to draft (pitch as tertiary)
**Project:** [`ADORSYS-GIS/cloud-identity-wallet`](https://github.com/ADORSYS-GIS/cloud-identity-wallet) + fleet patterns

### Hook / one-liner
Identity infrastructure must survive individual instance crashes without losing user work — the cloud wallet is designed so another instance can take over. This post shows the Rust patterns adorsys uses: async task flows, relocatable/retryable work, externalized encrypted state + wrapped keys, and circuit breakers around downstream issuers/verifiers/KMS in a distributed fleet.

### Why this post
Directly matches the "crash one instance, don't lose user work" requirement and the circuit-breaker work for robust decentralized servers. High differentiation.

### Content hints
- "Crash one instance, don't lose the user's work": design states another instance can claim and complete (handoff/idempotency).
- `tokio` task graphs, bounded retry with backoff, idempotent handlers.
- **Circuit-breaker** pattern for calls to decentralized parties (issuer/verifier/KMS) — fail fast rather than cascade.
- Honesty guardrail: frame as target failover/reproducibility **architecture** being implemented incrementally, not already-shipped claims.

### Acceptance criteria
- [ ] ~1000–1500 words, original
- [ ] ≥1 resilience pattern explained with reasoning
- [ ] Explicitly honest about shipped vs. in-progress
- [ ] Author bio + repo links included

_Part of epic #515 (#RF-BLOG-01)._