performance-optimization

Use a PerformanceCounter class with IDisposable timers to measure operation durations and counts, combined with request middleware that logs per-endpoint response times. Memory monitors using GC.GetTotalMemory track allocation patterns and collection counts.

Combine IMemoryCache as a fast L1 cache with IDistributedCache as an L2 fallback. Check memory first, then distributed cache, and populate memory on L2 hits to reduce serialization overhead on subsequent requests.

Cache-aside checks the cache before loading data from the source, then stores the result in cache on a miss. It is implemented with a generic service wrapping ICacheService and a data loader delegate, with explicit invalidation on writes.

Yes, EF.CompileAsyncQuery creates precompiled query delegates that skip repeated query translation, reducing overhead for frequently executed queries. Combine with AsNoTracking for read-only scenarios to further cut change-tracking costs.

High memory usage often comes from excessive allocations, unbounded collections, or missing object reuse. Use object pooling for frequently created objects like StringBuilder, stream large datasets with IAsyncEnumerable, and monitor GC collection counts.

Use batching when processing large item sets to avoid overwhelming downstream services or exhausting connection pools. Chunk items into groups, await each batch with Task.WhenAll, and add small delays between batches to throttle load.