Statistics

Here you can see statistics about the JVM (Java Virtual Machine) settings and parameters as well as the computation that is currently being processed.

Updating every ? second(s).

Maximum Graph Size Estimation

Here we try to estimate how large a data set might be at most to still be processable by your current infrastructure. This estimation, however, is quite vague and should not be taken for granted.

Logs

These are logging messages the system emitted while processing your data set. You can filter the messages according to their source and log level by using the following buttons:

Computation does not seem to start

When a computation does not seem to start, it might have crashed. The log messages below might indicate such a problem. However, as long as the charts below are still updating, a crash is unlikely. In such cases, the graph might take a lot of time to load. You can observe the graph loading stage by having a look at the charts regarding the infrastructure.

Computation does not converge

When a computation does not converge, there might be cyclic signal and collect messages so that it might never converge. You can observe the charts below to see whether there are any messages sent or received. We also show CPU time and load to investigate whether the node(s) is calculating anything. Additionally, you could observe the graph with a state visualization.

Computation crashes

A computation can crash for various reasons. There might be a problem in the the code (e.g. a NullPointerException) in which case you might see an error message in the list of log messages below. Another reason for a crash could be an infrastructure that is too weak to process the current data set. In this case, you might see an error message as well but could also observe the memory usage in the charts below. A further reason for a crashed computation might by that an OutOfMemory exception occurred. You may observe that by looking at the "Max Memory" and "Total Memory" charts below.

Computation takes very long

A computation on a large data set, or even a complex computation on a small data set, might take hours, days or even longer to compute. Please be patient. However, there might be other reasons that a computation takes very long. First, not every computation converges at all. Second, a high CPU usage shows that the computation is still running and you have to wait for it to end.

Computation yields unexpected results

There can be various reasons for when a computation yields unexpected results. To find potential problems, you can set breakpoints on the graph and step through different phases of the computation. You can also view the graph with state visualization enabled.

Node Charts

Here you can view all node charts that are generated and updated while processing your data. The different colors of the lines represent the minimum, average, and maximum values over all nodes. You can hover the lines to get more information.

Worker Charts

Here you can view all worker charts that are generated and updated while processing your data. The different colors of the lines represent the minimum, average, and maximum values over all workers. You can hover the lines to get more information.