Posted on by

Production Checklist for a Connected Project

Monochrome technical illustration for Production Checklist for a Connected Project

A prototype proves that a request can work. Production means it keeps working safely when credentials leak, networks fail, inputs are malformed, dependencies change, or the device needs recovery.

Minimum controls

  • Use an app API key with the narrowest scopes for each device.
  • Keep tokens in environment variables or a secret manager.
  • Use HTTPS, validate every input, and preserve trailing slashes.
  • Set timeouts and bounded exponential backoff for 429 and transient 5xx responses.
  • Log status, timing and request IDs without logging secrets.
  • Back up configuration and test restoration.

Retry and idempotency

Do not blindly retry an uncertain POST: it may create duplicates. Retry validation failures never; retry rate limits and transient server failures slowly and with a maximum attempt count. Use a client-side event ID or reconciliation process where duplicates matter.

A pre-release test run

  1. Use a non-production endpoint and a test credential.
  2. Send one valid record and confirm the expected 201 response.
  3. Send an invalid value and confirm the client receives a useful 400 without a record being created.
  4. Temporarily block the network and confirm the collector times out and recovers.
  5. Restart the service and verify it resumes without duplicating an uncertain write.
  6. Revoke the test key and confirm the client reports the authentication failure without exposing the key.

Device health and recovery

Run the collector under a service manager, verify restart behavior, monitor disk and memory, and keep a tested replacement image or setup procedure. Plan how to rotate a compromised key and disable a lost device.

Release review

  1. Test the current project and dependency versions.
  2. Confirm endpoint visibility, scopes and credential storage.
  3. Exercise sensor, network, API and dashboard failure paths.
  4. Record the rollback or recovery step.
  5. Review the checklist again after material changes.

Next step

Use REST API Authentication and Common Errors to diagnose a failed request, then apply this checklist before exposing the project to production traffic.