Skip to main content

​
Establish the deployment contract

Record the SDK version, device models, OS versions, transports, topology, enrollment policy and expected workload. Define retention, maximum offline duration, queue capacity and the application’s acceptance rules.

​
Preserve identity and state

Keep the same application/profile namespace across normal restarts. Store MLS material and protocol state through the platform’s supported providers. Treat logout, reinstall, key loss and device replacement as explicit lifecycle operations. Do not delete state as a general troubleshooting step. On Android, process restart requires an application-owned receiving path. Review registerMeshWakeTask and its manifest opt-in in the React Native integration guide. Register handlers before starting the protocol, including in background work.

​
Test the failure boundaries

  • Internet disconnected while peers continue communicating.
  • Peer partition, local work and later reconnection.
  • Process termination and restart with pending records.
  • Queue expiry and capacity exhaustion.
  • Rejected or duplicate business operations.
  • Backend outage, recovery and acceptance reconciliation.
  • Provider disappearance during a request, including uncertain outcomes.
Capture measured results for your deployment. Do not apply a simulator or local-LAN result to another radio or topology.

​
Operate the deployment

Provide actionable errors, pending count, oldest-work age and a support route. Hosted telemetry is optional; a local workflow must not depend on uploading analytics to operate. Pin versions, document the supported upgrade path and rehearse recovery before expanding the fleet. Read Upgrades and Licensing and hosted services. Optional telemetry is metered. Its client-side key can be extracted from the application binary; scope it to Mesh SDK, monitor usage and revoke it in the portal when needed.