2013 June 12 OpenStack

OpenStack Ceilometer数据与API源码详解

MongoDB的Collections

Ceilometer在MongoDB中共有这么几个Collections

- user
  - { _id: user id
      source: [ array of source ids reporting for the user ]
      }
- project
  - { _id: project id
      source: [ array of source ids reporting for the project ]
      }
- meter
  - the raw incoming data
- resource
  - the metadata for resources
  - { _id: uuid of resource,
      metadata: metadata dictionaries
      user_id: uuid
      project_id: uuid
      meter: [ array of {counter_name: string, counter_type: string,
                         counter_unit: string} ]
    }

其中meter是采集到的数据，其他的都是固定值

Collector对数据库的写数据

Collector在接收到采集的数据后，会调用record_metering_data()对数据进行写入，相应mongodb的代码在ceilometer.storage.impl_mongodb中

def record_metering_data(self, data):
    self.db.user.update(
        {'_id': data['user_id']},
        {'$addToSet': {'source': data['source'],},},
        upsert=True,
    )
    self.db.project.update(
        {'_id': data['project_id']},
        {'$addToSet': {'source': data['source'],},},
        upsert=True,
    )

    self.db.resource.update(
        {'_id': data['resource_id']},
        {'$set': {'project_id': data['project_id'],
                  'user_id': data['user_id'],
                  'metadata': data['resource_metadata'],
                  'source': data['source'],},
         '$addToSet': {'meter': {'counter_name': data['counter_name'],
                                 'counter_type': data['counter_type'],
                                 'counter_unit': data['counter_unit'],},},
         },
        upsert=True,
    )

    record = copy.copy(data)
    self.db.meter.insert(record)
    return

从上面代码可知，每次存储时都会更新user,project和resource，然后将数据完全写入到meter中，写入后的数据格式如下：

{
  "counter_name": "disk.write.requests",
  "user_id": "4ff44f4665564b2abcb8e1f1619f2b85",
  "message_signature": "8473976666aecd078a281afed936839b737ceaf4bb63654759d63514bdc9ee03",
  "timestamp": ISODate("2013-05-21T22:33:14.0Z"),
  "resource_id": "b7fc623d-1d4a-4ac7-b96b-78c9d921fa74",
  "resource_metadata": {
    "ramdisk_id": "",
    "display_name": "test",
    "name": "instance-00000001",
    "disk_gb": "",
    "availability_zone": "",
    "kernel_id": "",
    "ephemeral_gb": "",
    "host": "e781ff9ce97dcc328d8826cfb19a20c001b866cb20859653c2f481b1",
    "memory_mb": "",
    "instance_type": "42",
    "vcpus": "",
    "root_gb": "",
    "image_ref": "da04e6dd-4cc7-4594-87d8-60927c07c396",
    "architecture": "",
    "os_type": "",
    "reservation_id": "",
    "image_ref_url": "http:\/\/192.168.0.6:8774\/676730085ab84296a9b4a7d68ee76078\/images\/da04e6dd-4cc7-4594-87d8-60927c07c396"
  },
  "source": "openstack",
  "counter_unit": "request",
  "counter_volume": NumberInt(1366),
  "project_id": "be13e080970d44b280e4843e084bb2b1",
  "message_id": "6cf1d76c-c266-11e2-a987-5eafb2e29593",
  "counter_type": "cumulative"
}

这是一个disk.write.requests的数据，其中resource_metadata如果无变化的话，没个都会带这些数据，具体原因不详

另外，重要的东西在

counter_unit 计量单位
counter_volume 计量数值
counter_type 计量类型

计量内容

在文档中讲了计量值和其单位

首先是计量类型：

Cumulative 随时间的累计值（如cpu总时长）
Gauge 离散项(floating IPs, image uploads)和变化的值 (disk I/O)
Delta 随时间的变化量（带宽等）

计量单位比较多了，每个都不太一样，这个可以查询文档，如磁盘读写请求的单位为”request”

API对数据库的读操作

数据存储只是Ceilometer的一小部分，如果合理的利用和分析采集到的数据才比较重要，另外这部分也是暴露出来给开发者的部分

API服务以wsgi service方式运行在后端，Ceilometer有v1和v2两个版本的API，v1会被弃用，这里只讲v2部分

GET /v2/meters/cpu?q.op=ge&q.op=lt&q.op=eq&q.value=2013-05-19+23%3A00%3A00&q.value=2013-05-20+00%3A00%3A00&q.value=b7fc623d-1d4a-4ac7-b96b-78c9d921fa74&q.field=timestamp&q.field=timestamp&q.field=resource

这是一个我截取下来的请求，首先我们看到一个资源地址

GET /v2/meters/cpu

根据V2的Controller，我们可以看到是MetersController()对它进行处理的

class V2Controller(object):
    resources = ResourcesController()
    meters = MetersController()
    alarms = AlarmsController()