@Nuvei-Robar,
Communication latency is usually very small, I would say in almost all cases below 100ms.
But I believe you are asking for the time between when a metric was sent to Dynatrace (by OneAgent,API, etc.) and when it appears somewhere in Dynatrace (eg. Data Explorer, Dashboards, Alerts etc.).
The following info is based on my experience, and might not correspond to reality. Other might complement/correct me.
I believe the major factor for the latency involves what a metric really represents in Dynatrace. Most metrics that I work with have a one minute resolution. So, I believe Dynatrace waits for that full minute to complete, before the metric point is available. There should be a lot of exceptions here, as you typically can see trace level metrics before that, for instance.
If you send a metric point at like 00:00:00.100, that metric will belong to the 00:00-0:01 minute "bucket". Till 00:00:59.999, if you send more data for that metric, it will be consolidated in a new value. And after that, Dynatrace will take some time to show you the value.
It gets even more trickier when you consider that you can inject data until one hour in the past, and 10 minutes in the future. How does Dynatrace recalculate values is something that I don't know, or have tested.
If you are injecting data like in a crontab sequence, you might be grabbing data from the last minute. In this case, if going through an extension or API, I recommend you timestamp the data as belonging in the last minute, which of course makes more sense, and gives a more faster reaction.
It seems that there might also be great differences by how this is dealt in Grail, or in Cassandra/ElasticSearch in Managed.
I believe the following Community posts also help:
Antonio Sousa
