why would a skipped policy take >1 ms?

[akoo]

why would a skipped policy take >1 ms? For most API calls containing skipped policies, the skipped policies take 1ms or less. Would shared infrastructure and load affect this situation of skipping a policy?

Key assumption: The condition on the skipped policy is does not seem to have bearing on the time. The majority of times the policy is run, the time <1 ms.

[adas]

@Alex Koo Very good question. Are you seeing this from trace ? At times I have seen trace to be a little unreliable depending on where exactly in the flow you are, whether you are calling a target or null target or node.js script target etc. Not saying this is a bug with the trace UI, but the debug session data that gets captured has an issue at times. Ideally skipped policies should not take more than 1 ms, but you are right, if there are scenarios where the runtime is under stress certain conditional evaluations may take a little longer. Have you tried this on a runtime (message-processor) that is not shared or experiencing huge amount of load/stress ?

[akoo]

Indeed, this is in Trace UI. I didn't consider the UI as being a factor, but I will investigate further. Lately, I have also seen this behavior on non-shared MPs. but it's not all of the time. I would not think C* being shared as having any impact on skipped policies. Thanks for the reply, @arghya das.

[#4]

Usually ,skipping a policy would depend on a prior condition (ie we would have a conditional statement to be checked).I thought the 1ms was the time it took to check the condition and then skip the policy.Makes sense ?