Understand Confluent Control Center

Question

Can someone help me understand what the red stripes on the time line mean in confluent control center?

Answer 1

The red stripes indicate that either missing or duplicated interceptor records were encountered when processing the monitoring data. The docs describe this behavior as such:

It’s also possible for messages sent by the Confluent metrics interceptors to be lost or duplicated. If this occurs, the affected time range is highlighted by showing a herringbone pattern on the axis.

You may find the Stream Monitoring section of the Control Center documentation helpful as you begin to familiarize yourself with the product. I have linked this below.

http://docs.confluent.io/current/control-center/docs/monitoring.html

Answer 2

The "red stripes" means Control Center can calculate message counts and lag, but there is uncertainty about the measurements.

The reason there is uncertainty about the measurements is that some of the information used to determine is gone missing. Typically the missing information is the "end of session" markers from the producers or consumers, and the reason it is missing is that the client shut done uncleanly (without calling close() method). Make sure you are cleaning up after your client even if exceptions are thrown and ctrl-c is hit.

In theory, the area of uncertainty should be pretty small. Just few minutes on both sides of the unclean shutdown. However, there is a bug in existing releases of Control Center, up to and including 3.3.0 that causes the "herringbone pattern" (our name for the red stripes) to take over the entire timeline. We are working on fixing this.