while our data is "real-time" in the sense that we process events as they come in, there's still a lot of processing involved in the calculation of a seemingly simple metric as "Response Time" etc.
First, the agents need to send the events that will eventually lead to a purepath on the server. These events have to pass over the network via our collectors. Events from many different agents can make up one purepath. The server needs to wait for all these events and correlate them into the purepath. Once this purepath is complete, it's passed along to the component that calculates the metrics. All this can be done in a lot less than one minute, but often there's a lot of traffic and the queues get filled up and there's a slight delay.
I personally wouldn't be concerned about a one minute delay, but if you want your measures to be as current as possible, make sure that the queues never fill up, i.e. that the sizing is adequate for the amount of data coming in.
Inspect the self monitoring dashlets to this purpose (Deployment Health -> Server Health etc) and make sure that the network's as fast as possible.
Note that these were general remarks without taking into account potential errors in your specific case, so if you feel that my explanations are insufficient, please open a support ticket with support archive and session so that we can help you further. Thanks,
Has the load increased (e.g. more agents / more users of your app)? Has there been a network modification? It's not really possible to give a detailed reason for this without analyzing logs and performance metrics.
If this is critical for you, please open a support ticket and make sure to include a self-mon session with the support archive (should be included automatically). You may want to reference this forum discussion there.