A Java Lambda that generates rare monthly utility reports shows multi-second cold starts when the first citizen clicks after long idle periods. Idle cost is a concern for infrequent use. What is the most appropriate performance approach?
Select an answer to reveal the explanation.
Short Explanation
Cold starts on a sleepy monthly report feel like waiting for a library printer to wake up. Fix the package and runtime first; buy provisioned concurrency only if the SLA truly needs always-warm capacity. Zero reserved concurrency just blocks everyone.
Full Explanation
Lambda cold starts are influenced by runtime, dependency size, and initialization work. For infrequent workloads, shrink the deployment artifact and consider a lighter runtime before paying for provisioned concurrency; use provisioned concurrency when latency SLOs justify the idle cost. Setting reserved concurrency to zero throttles invokes, and dashboards alone cannot replace compute for report generation.