9.6.4.5 Understanding the Metric Stream Format

Real-time metric observations contain a core set of attributes. However, the format of the metric stream depends on the mode of access.

If you are automatically uploading the metric stream to a metric collection platform, you can specify the metric stream format in the metric endpoint configuration. The available formats are JSON or plain text.

Following is an example of the JSON format:

json: {
    "gauge": [{
            "metric": "OS_NET_RX_BY_SEC",
            "value": "0.0012989044189453125",
            "timestamp": 1652473286000,
            "unit": "MB/sec",
            "dimensions": {
                "server": "celadm09.example.com",
                "objectName": "eth0",
                "nodeType": "STORAGE",
                "fleet": "example-fleet",
                "pod": "dbm01",
                "cluster": "c01"
            }
        }, {
            "metric": "SIO_IO_RD_FC_HD_SEC",
            "value": "0.0",
            "timestamp": 1652473286000,
            "unit": "MB/sec",
            "dimensions": {
                "server": "celadm09.example.com",
                "objectName": "SMARTIO",
                "nodeType": "STORAGE",
                "fleet": "example-fleet",
                "pod": "dbm01",
                "cluster": "c01"
            }
        }
    ]
}

The plain text format contains essentially the same information as the JSON stream. However, with the plain text format, each metric observation is presented on a separate line. Following is an example of the plain text format:

metrics,name=OS_NET_RX_BY_SEC,objectName=eth0,server=celadm09.example.com,unit=MB/sec,nodeType=STORAGE,fleet=example-fleet,pod=dbm01,cluster=c01 value=9.441184615324398E-4 1652473456000000000
metrics,name=OS_NET_RX_BY_SEC,objectName=eth0,server=celadm09.example.com,unit=MB/sec,nodeType=STORAGE,fleet=example-fleet,pod=dbm01,cluster=c01 value=0.002647613311980988 1652473457000000000

If you are downloading the metric stream by using the provided REST endpoint, the data is presented in a format similar to the plan text upload format where each metric observation is presented on a separate line. Following is an example of the download format:

DS_CPUT{objectName="dbadm05",unit="%",server="dbadm05.example.com",nodeType="KVMHOST",fleet="example-fleet",pod="dbm01",cluster="c01"} 23.10906363831155 1652485449597
DS_MEMUT{objectName="dbadm05",unit="%",server="dbadm05.example.com",nodeType="KVMHOST",fleet="example-fleet",pod="dbm01",cluster="c01"} 99 1652485449597
DS_MEMUT_MS{objectName="dbadm05",unit="%",server="dbadm05.example.com",nodeType="KVMHOST",fleet="example-fleet",pod="dbm01",cluster="c01"} 0.12396045794483294 1652485449597

The following list describes the attributes contained in the metric stream:

• The metric name is identified as follows:

  • In the JSON upload format, the metric name follows the metric tag.

  • In the plain text upload format, the metric name is the value following name=.

  • In the download format, the metric name is the first element on each line, preceding the left braces ({).

• The metric value is located as follows:

  • In the JSON upload format, the metric value follows the value tag.

  • In the plain text upload format, the metric value follows value=.

  • In the download format, the metric value is the second last element on each line, following the right braces (}).

• The time of the metric observation is located as follows:

  • In the JSON upload format, the timestamp follows the timestamp tag. The timestamp is expressed as the number of milliseconds (1 x 10^-3 sec) since January 1, 1970, 00:00:00 GMT.

  • In the plain text upload format, the timestamp is the last element on each line. The timestamp is expressed as the number of nanoseconds (1 x 10^-9 sec) since January 1, 1970, 00:00:00 GMT.

  • In the download format, the timestamp is the last element on each line. The timestamp is expressed as the number of milliseconds (1 x 10^-3 sec) since January 1, 1970, 00:00:00 GMT.

• The unit value describes the unit of measurement for the metric observation.

• The server value contains the name of the Exadata server that generated the metric observation.

• The objectName value contains the name of the Exadata object associated with the metric.

• The nodeType value contains the type of the Exadata server that generated the metric observation.

• The fleet, pod, and cluster attributes are examples of user-defined metric tags, which you can use to organize and group observations generated by numerous Exadata servers. You can tag metrics by setting the metricStreamTags DBSERVER attribute.