The State Changers meeting revolved around the discussion of using lambdas in Xano. The participant asked a basic question about improving performance while reducing cost, particularly for small endpoints that are triggered frequently. It was clarified that performance in Xano is largely proportional to the number of instances of given queries. While JavaScript Lambdas may not require higher load, the time a process takes can impact the overall server load either by increasing or decreasing it.
The theoretical possibility of Lambda becoming inefficient was also discussed: if a Docker container implies a large number of launch instances, this would bring inefficiencies due to the 'cold start' problem. It was suggested that for peak performance tweaks, lambdas should not be put inside loops, but rather to marshal the information separately and give it to the lambda, iterating over the lambda's return to keep it out of the loop. Overall, identifying bottlenecks and addressing them was seen as a more effective optimization strategy. This would potentially involve a deep look into request history, prioritizing ones that are frequent and slow. Direct query optimization such as adding an index or avoiding queries inside loops can often unlock improved performance system-wide. The discussion emphasized that in a functioning system, there is always an element acting as the bottleneck to address. The idea of doing a deep dive on the subject of performance was proposed for future meetings.
(Source: Office Hours 8/2/2023 )
