Calling a Model Is Easy. Running an AI System Is Not
Why a successful model call says little about whether an AI feature is reliable, secure, observable, affordable, and ready to operate in production.
Reliability
Choose rate-limiting policies by endpoint cost
Apply different ASP.NET Core rate limits to cheap reads and expensive AI or data-processing endpoints.
Design queue handlers for duplicate delivery
Assume a queue message can arrive more than once and make the handler’s business effects safe to repeat.
Validate structured model output before using it
Treat deserialized model output as untrusted data and apply normal business validation before it drives application behavior.
Use PeriodicTimer for cancellable asynchronous loops
Use PeriodicTimer when a background loop needs an awaitable cadence, cooperative shutdown, and sequential executions.
Validate required options at startup
Fail during application startup when required configuration is missing or invalid instead of waiting for the first request.
Use idempotency keys for retryable writes
Make a retried write replay the same operation instead of creating a duplicate side effect.
Make logs structured before adding dashboards
Capture queryable fields first so dashboards are built on data, not text search.
Use IHttpClientFactory
Create configured HttpClient instances through the factory instead of scattering clients and handlers.
Do not hide exceptions inside fire-and-forget tasks
If work can fail, give it ownership, logging, cancellation, and a lifecycle.