Reliability & Monitoring
Smart wallet infrastructure has one job: be there when a transaction needs to go through. A stuck transaction isn't an inconvenience, it's a stuck user, a failed payment, a support ticket you can't explain.
How we verify liveness
Every minute, on every chain we support, we execute a real signed transaction end to end. This confirms the full path works: RPC connectivity, bundler availability, and finality. A node health check only confirms a server responded, it says nothing about whether a transaction actually landed.
RPC redundancy
Every chain we support runs behind multiple independent RPC providers. If one degrades or goes down, traffic fails over automatically, without operator intervention and without a visible interruption to requests in flight.
Monitoring built for chains
Chain failures look different from server failures: stuck mempools, reorgs, gas spikes, nonce gaps. We built monitoring for these specifically. An issue triggers an alert before it reaches a production wallet.
SLA by plan
| Plan | SLA |
|---|---|
| Starter | No SLA |
| Launch Prod | No SLA |
| Grow | 99% uptime |
| Enterprise | 99.9% uptime, or as specified in contract |
Full terms, downtime definition, and service credits are in our Terms of Service.
Live status
Current and historical uptime is public at status.candide.dev.