Steve Nyemba f51a870fd8 bug fix: ... | 1 year ago | |
---|---|---|
.idea | 7 years ago | |
bin | 1 year ago | |
smart | 1 year ago | |
test | 6 years ago | |
.gitignore | 3 years ago | |
.gitmodules | 7 years ago | |
README.md | 6 years ago | |
info.py | 1 year ago | |
readme.md | 3 years ago | |
setup.py | 1 year ago | |
start.sh | 7 years ago |
This program is designed to monitor processes running on a linux/osx machine. It assumes logging hasn't been properly done. The program answers basic questions:
- Is a given program still running
- How much resource (memory/cpu) a program is using up
- The number of processes found
- Folder monitoring ...
#Architecture
The architecture of the system is distributed with a central master node,
{
"id":"",
"key":"",
"apps":[],
"sandbox":[{"path":"","requirements":""}],
"folders":["path-1"],
"store":{}
"actions":{}
}
The agent will perform three basic functions :
- data collection (building up the dataset)
- classification of status for a given process:
x y z
0 0 1 idle
1 1 1 running
0 0 0 crash
with x : memory used
y : cpu usage
z : number of processes found
- Prediction of crashes using multivariate anomaly detection
This allows the learner to determine if something unusual is happening.