Send Data

You can send logs, events, and metrics to the Cloud Monitoring service. To send data to the service, use a separate client class - SendClient. Create an instance of this class:

MonitoringServiceClient.send()

To send logs, use the logs method:

MonitoringServiceClient.send().logs(source, logs);

The log record format is as follows:

{

"objectType":"log", // required

"timestamp":"2018-07-31T11:07:02.312", // required; formats: [ISO8601, UNIX_MS]

"message":"Hello, Monitoring!", // required; plain text

"severity":"DEBUG", // optional; plain text

"parameters":{ // optional; you can use any names for sub-fields

"logger":"xes.system", // optional

…

}

}

Note: All unspecified root-level fields are dropped.

To create a log record, use the LogRecord class with a chained (or fluent) builder:

new LogRecord()

.withMessage("some log message 2")

.withTimestamp(System.currentTimeMillis())

.withParameters(new Parameters()

.put("key", "value3")

.put("key2", "value4"))

You can use the following parameters:

  • any Map<String, Object>
  • instance of Parameters
  • any object (Jackson ObjectMapper is used)

To send metrics, use the metrics method:

MonitoringServiceClient.send().metrics(source, metric);

The metric record format is as follows:

{

"objectType":"metric", // required

"timestamp":1533065825066, // required; formats: [ISO8601, UNIX_MS]

"Memory Usage (Kb)": { // required; an object cannot have the “objectType”, “timestamp” and “parameters” names.

// Can contain one or more sub-fields with arbitrary names and numeric values

"AVERAGE":107696.0, // required; numeric [int, long, float, double]

"MIN":67567, // optional; numeric [int, long, float, double]

"MAX":147825, // optional; numeric [int, long, float, double]

"COUNT":2, // optional; numeric [int, long, float, double]

"SUM":215392, // optional; numeric [int, long, float, double]

"LATEST":67567 // optional; numeric [int, long, float, double]

},

"parameters":{ // optional; any names can be used for sub-fields

"name":"profile=Default", // optional

"type":"profile", // optional

...

}

}

To create a metric record, use the MetricRecord class similar to the LogRecord class.

To send events, use the events method:

MonitoringServiceClient.send().events(source, events);

The event record format is as follows:

{

"objectType": "event", // required

"timestamp": "2018-07-31T12:27:02.312", // require;, formats [ISO8601, UNIX_MS]

"name": "XEServer is STOPPED", // required; plain text

"parameters": { // optional; you can use any names for sub-fields

"system":"/", // optional

"type":"XES Command Event", // optional

...

}

}

Note: All unspecified root-level fields are dropped.

To create an event record, use the EventRecord class similar to the LogRecord class.

You can send logs, events, and metrics as:

  • binary data stream

    • SendResult logs(String source, InputStream data)

      Parameters:

      source: a source name validated using the following regular expression: [\w+\-_!.'()]{1,255}

      data: binary data stream in the utf-8 format. Has to be a valid JSON string.

  • binary data

    • SendResult logs(String source, byte[] data)

      Parameters:

      source: a source name validated using the following regular expression: [\w+\-_!.'()]{1,255}

      data: binary data in the utf-8 format. Has to be a valid JSON string.

  • records

    • SendResult logs(String source, LogRecord... records)

      Parameters:

      source: a source name validated using the following regular expression: [\w+\-_!.'()]{1,255}

      data: an array of LogRecord

  • records from a collection of objects. You can use LogRecord, EventRecord, MetricRecord, or your custom object. The system uses a Jackson mapper.

    • SendResult logs(String source, Collection<LogRecord> records)

      Parameters:

      source: a source name validated using the following regular expression: [\w+\-_!.'()]{1,255}

      data: an array of LogRecord

Note: We recommend you send data to the Cloud Monitoring service as a binary data stream using the InputStream methods. When you do this, data is sent in the HTTP stream that does not require data buffering.