Serverless and Cloud Cost-Performance Review
Balance latency and cost in cloud functions.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
[VERIFY] is an editorial flag, not a blank to fill. Wherever the prompt shows [VERIFY], check that item against a reliable source before you use it.
Review my cloud functions for latency and cost. Provider and runtime: [PROVIDER_AND_RUNTIME]. Function data: name, memory setting, average and p95 duration, invocation count per month, cold start frequency, timeout, and cost if known: [FUNCTION_DATA]. Architecture notes: [ARCHITECTURE_NOTES]. Output as tables: 1. Function table: Function | Monthly invocations | Memory | Avg duration | Estimated monthly cost (show the formula using the provider's pricing model; mark price inputs as [VERIFY]) | Issue flagged. 2. Issue table: Function | Problem (over-allocated memory, cold starts, long timeouts, chatty calls to a database, large package size, synchronous work that could be queued) | Evidence | Fix | Expected effect. 3. Memory tuning experiment plan: Function | Test memory sizes | Metric to compare | Decision rule (choose the cheapest setting that meets the latency goal). 4. Architecture options: keep as functions, batch, queue, or move to a container, with the break-even condition for each. Finish with the three changes with the largest saving and the risks of each. Do not assume prices; show the formula so I can enter current ones. Verify every fact, figure and claim in the output against your own records or the official documentation, and have a qualified professional check anything with legal, safety or financial consequences before you rely on or publish it.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.