Okay, so one proxy flow performs a service call out to another flow/path of the same proxy. At the PreFlow stage of the proxy I have a SpikeArrest policy set at 10ps. This one triggers a 429 (too many requests).
I remember having read that SpikeArrest interprets Nps as N segments of a second - each segment 1 Nth of a second long. And only exactly one request per segment is accepted.
Obviously (and hopefully) a service callout will perform its request very fast - so it falls into the same segment as the initial call to the proxy.
What is the recommended approach here?
10ps means 1 request in every 100 miliseconds. You can see the time elapsed from the beginning to the service callout. If that is within 100 miliseconds oviously you will get this error. You need to increase the spike arrest limit else you can use identifieer with flow name or path
> 10ps means 1 request in every 100 miliseconds. You can see the time elapsed from the beginning to the service callout. If that is within 100 miliseconds oviously you will get this error.
that's what I'm explaining in my problem description, aren't I ...
> You need to increase the spike arrest limit
certainly not ... that's not a solution.
> you can use identifieer with flow name or path
not sure what you mean.