We received reports that the timeout for the JVM tests was too low, so I rewrote the MetricsReader class. Now it uses commons-io to follow the log close to live (polling once a second) and stops when it finds the right metrics. As a result we can safely increase the maximum timeout significantly.
We received reports that the timeout for the JVM tests was too low, so I rewrote the MetricsReader class. Now it uses commons-io to follow the log close to live (polling once a second) and stops when it finds the right metrics. As a result we can safely increase the maximum timeout significantly.