Failure Handle should be improved to provide more information about failures based on existing requests and experince.

The main concern is the reporting and ability to find information. The failure handler should have a summary
with failure analysis based on .jtr report and/or tools output

The possible known patterns:
1. Ability to get jcmd output/complete jcmd commands.
2. Ability to run native tools.
3. dmesg output containigs OOM Killer messages
3. Errors in SA output?
4. Just summarize failure sings


The another part is support of java and native thread stacks. (Covered by separate bug)
1. Color syntax of output.
2. Comparison of stack output for different times.

The thread analysis might also be part of summary.

The Hotspot might be improved to generate more data if test times out because of VM problems.
If VM fails to enter/exit safepoint it shoul print somehow diagnostic information.

Other ideas that might require more separate tasks:
The symbols should be unpacked and used during analysis, see https://bugs.openjdk.org/browse/JDK-8269256.
Also, hsdis might be used to generate more info.
The jfr, rr files might be processed to some text format for fast analysis.