通过进程id找到进程对应的容器并统计每个进程的内存占用写到excel里

 # coding=utf-8
 import re
 import os
 import commands
 import json
 import psutil
 from pyExcelerator import *

 def execute(cmd):
     status, output = commands.getstatusoutput(cmd)
     if status != 0:
         raise Exception('status is %s, output is %s' % (status, output))
     return output

 def get_all_container_ids_name():
     infos = execute("docker ps |awk '{print $1, $NF}'").split('\n')
     all_ids = {}
     regex = re.compile('\s+')
     for info in infos:
         docker_id, docker_name = regex.split(info)
         short_id = docker_id.strip()
         if short_id.strip().startswith('CON'):
             continue
         full_id = execute("docker inspect -f '{{.Id}}' %s" % short_id)
         state = execute("cat /run/runc/%s/state.json" % full_id)
         f = json.loads(state)
         cgroup_paths = f['cgroup_paths']['pids']
         pids_path = os.path.join(cgroup_paths, 'cgroup.procs')
         ids = execute("cat %s" % pids_path).split('\n')
         for prgress_id in ids:
             pr_id = prgress_id.strip()
             all_ids[pr_id] = {'id': short_id, 'name': docker_name}
     return all_ids

 def get_process_info(p):
     try:
         # cpu = int(p.cpu_percent(interval=1))
         rss = p.memory_info().rss
         name = p.name()
         pid = p.pid
         return '%s,%s,%s\n' % (pid, name, rss)
     except Exception as e:
         print e.message

 def get_all_process_info():
     """取出全部进程的进程名，进程ID，进程实际内存, 虚拟内存,CPU使用率
     """
     node_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'node_test.log')
     instances = ''
     all_processes = list(psutil.process_iter())
     for proc in all_processes:
         ret = get_process_info(proc)
         if ret:
             instances += ret
     with open(node_name, 'w') as fp:
         fp.writelines(instances)

 def get_docker_name():
     file_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'node_test.log')
     result_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data.txt')
     id_name_relation = get_all_container_ids_name()
     tmp = ''
     regex = re.compile(',')
     with open(file_name, 'r') as fp:
         for progress_info in fp.readlines():
             progress_id, mem, progress_name = regex.split(progress_info)[0], regex.split(progress_info)[2], \
                                               regex.split(progress_info)[1]
             if progress_id in id_name_relation:
                 tmp += '%s %s %s %s %s\n' % (progress_id, id_name_relation[progress_id]['id'],
                                              id_name_relation[progress_id]['name'], progress_name, mem)
             else:
                 tmp += '%s %s %s %s %s\n' % (progress_id, 'sys_progress', 'sys_progress', progress_name, mem)
     with open(result_name, 'w') as fp:
         fp.writelines(tmp)

 def ge_excel():
     file_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data.txt')
     result_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data.xlsx')
     regex = re.compile(' ')
     w = Workbook()  # 创建一个工作簿
     ws = w.add_sheet('node_1_data')  # 创建一个工作表
     ws.write(0, 0, 'pid')
     ws.write(0, 1, 'docker_id')
     ws.write(0, 2, 'docker_name')
     ws.write(0, 3, 'progress_name')
     ws.write(0, 4, 'mem(MB)')
     index = 1
     with open(file_name, 'r') as fp:
         for info in fp.readlines():
             progress_info = info.strip()
             if progress_info:
                 progress_id, docker_id, docker_name, progress_name, mem = regex.split(progress_info)
                 ws.write(index, 0, progress_id)
                 ws.write(index, 1, docker_id)
                 ws.write(index, 2, docker_name)
                 ws.write(index, 3, progress_name)
                 ws.write(index, 4, float(mem) / 1024 / 1024)
                 index += 1
     w.save(result_name)

 def delete_tmp_file():
     data_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data.txt')
     node_test_name = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'node_test.log')
     if os.path.exists(data_name):
         os.remove(data_name)
     if os.path.exists(node_test_name):
         os.remove(node_test_name)

 if __name__ == '__main__':
     delete_tmp_file()
     get_all_process_info()
     get_docker_name()
     ge_excel()
     delete_tmp_file()