The report’s first unspoken insight: customers ask for fast RPC, but the deeper job is not having to debug why production died. A single uptime number is no longer enough when rate limits, degraded endpoints, archive/log workloads and chain-specific failures surface inside the application.
Product proposal
- Private Health API: per-chain and per-method status, 429/503 attribution, endpoint degradation states.
- RPC Failure Report: customer-specific diagnostics that explain what failed, where and why.
- RCA template: incident explanation and prevention actions after material degradation.
- Synthetic monitoring: customer-relevant methods tested by region and chain.
- Optional failover recommendations: keep GetBlock as the control plane instead of becoming one endpoint inside someone else’s router.
Validation
Run a dashboard pilot with 5–10 production customers. Track whether it reduces review/debug time, support tickets, incident anxiety and churn risk. Pair it with 10 win/loss interviews focused on latency vs price vs chain count vs incident tooling.