Join us in this series of weekly office hours, where you'll dive deep into the world of APIs and integration on Google Cloud with Apigee and Application Integration. Whether you're a seasoned Google Cloud pro, partner, customer, or just curious, this is your opportunity to learn & connect alongside experts.
We'll meet on Thursdays at 4:00 PM CET for engaging, hour-long sessions packed with insights - Don't miss out & register now!
Thank you to everyone who joined us! 👥 We appreciate your participation and engagement throughout the session. As promised, find below the recording and deck. We look forward to seeing you at our future events!
Hello everyone! We’re excited see you today! 🎉 If you haven’t already, please complete your official registration here, to ensure you receive the reminder email with the link and access details.
Feel free to share your questions in advance for the open discussion here - we look forward to seeing you there!
Nice office hours session!
Regarding observed latency (time of 37:50 onward), in general it's not a good idea to draw conclusions regarding performance based on one or two requests. A valid perf test needs to run over a period of time, with a typical request load.
In this scenario, if you run a single service Callout or ExternalCallout, and measure the performance of that, you may incur connection setup time for the first few requests (one setup for each MP-to-service pair). And you won't need to pay that cost for subsequent calls that re-use the same connection .Also there can be intervening network systems between those two pieces, that also need to "warm up". So I think the comparison of 2ms for in-process JavaScript vs 28ms for ExternalCallout is probably not a valid one. I doubt the majority of the cost for the 28ms is incurred in the Javascript logic! Just guessing though.
Yes indeed, testing is part of SDLC. It is also a good idea to raise all the factors, causes and hypothesis of low performance (latency) regarding External or Service Callout. Once factors, causes, hypothesis raised, test and find the right results to solve the low latency in external and service callout.