Success case

No internet

Server side issues with 500 error codes as response

User side request issues with 400 error codes as response

Cases where your network throwed a random exception like no host found

Timeout issues

Authentication failures

Retry failed requests or after re-auth

Handling null / unexpected values from response

Create user friendly error messages instead of throwing dev errors to users by parsing errors

Cancelling few requests when user goes background after calling it and few requests to run the same like analytics

making batched requests from DB

Retry with exponential backoff

Storing cache, send cache first for faster response and make actual network call + cache invalidation when new response shows up

Doing all of these when there are multiple serial + parallel background tasks / api calls running at same time.

Ensuring the transformations to create ui model from network model doesn’t create a lag in user device.

Adding observability traces to these calls

Debouncing, throttling accordingly

Handling app process deaths / crashes

unit testing , integration tests with mock / fake

load testing, test of low networks

ensuring everything is behind feature flag so app works same as before with some failures in new implementation