Notifications
Clear all
Topic starter
A customer of yours has an SLA with their client that a particular service will respond within 4 seconds. The end client has reported that it feels slower. Your engineers do a trial at the client site and notice that there seems to be a delay for many of the requests. It's your team's responsibility to identify the issue quickly within the strict timeline for fixes according to the contract, and then fix it .
What should you do?
- A . Recommend a move to serverless technologies which will scale automatically on demand.
B. Add logging statements at multiple points in the application, build it, and deploy it. Now new requests will give us information on latency in the logs.
C. Check if the browsers used by the client are different from yours. If they are, that's most likely the issue. Ensure that everybody uses the latest version of the browser that you are also using.
D. Use Cloud Trace to collect latency data and track how requests propagate and why there is a delay.
Suggested Answer: D
Explanation:
Cloud Trace is a built-in tool in the Operations suite to identify issues like latency.
-> Such fixes are unlikely to change core issues like the service itself being architected or written sub-optimally. Though changes like browser, networking, etc. are helpful, it would be the wrong approach to first recommend that the customer upgrade all their hardware and software.
-> Rewriting code and logging information is going to be time consuming. In general though, logging should always be included in code and it can give good insights. But tracing is way more specific and comprehensive for this requirement.
-> In certain cases, we might identify scaling as the issue. But we should first identify the core problem. So, start with tracing. We can also achieve scale in server-ful technologies.
Reference link- https://cloud.google.com/trace
Explanation:
Cloud Trace is a built-in tool in the Operations suite to identify issues like latency.
-> Such fixes are unlikely to change core issues like the service itself being architected or written sub-optimally. Though changes like browser, networking, etc. are helpful, it would be the wrong approach to first recommend that the customer upgrade all their hardware and software.
-> Rewriting code and logging information is going to be time consuming. In general though, logging should always be included in code and it can give good insights. But tracing is way more specific and comprehensive for this requirement.
-> In certain cases, we might identify scaling as the issue. But we should first identify the core problem. So, start with tracing. We can also achieve scale in server-ful technologies.
Reference link- https://cloud.google.com/trace
Posted : 18/12/2022 8:55 am