Articles by Ranknod

Build a test matrix for migrating .NET Azure Functions from the in-process model to the isolated worker model, covering triggers, HTTP contracts, deployment, telemetry, and rollback.

Design an idempotent C# webhook ingress using delivery IDs, durable inbox records, database uniqueness constraints, and a transactional outbox to prevent duplicate agent actions.

Learn how to prevent sensitive prompts, payloads, and exception details from reaching .NET AI application logs by using an explicit structured logging contract.

Build and test a reusable Polly circuit breaker in .NET 8 around an AI provider call, including open, half-open, failed probe, and recovery states.

Validate embedding dimensions, model compatibility, finite values, and zero vectors in C# before using vectors for cosine-similarity search.

Learn how to enforce tenant isolation in a .NET RAG application by deriving trusted tenant identity and filtering documents before ranking and returning results.

Build a C# token-budget ledger for multi-call AI workflows that reserves token allowances, reconciles reported usage, and blocks calls when the shared budget is exhausted or uncertain.

Learn how to control concurrent AI requests in .NET with bounded channels, fixed workers, cancellation handling, and explicit queue admission limits.

Learn how to distinguish caller cancellation from application timeouts in .NET AI requests using linked cancellation tokens, explicit outcome states, and cooperative cancellation.

Learn how to prevent duplicate articles when publishing API calls time out using stable idempotency keys, unique database jobs, retries, and reconciliation.

Learn how to design reliable AI agent workflows in C# using state machines, explicit transitions, approval states, retries, persistence, authorization, and execution controls.