Trace data formats - Simple Log Service - Alibaba Cloud Documentation Center

This topic describes the trace data formats that are supported by Simple Log Service.

Simple Log Service is compatible with the trace data formats that are defined in OpenTelemetry Trace 1.0. If trace data is written by using protocols such as OpenTelemetry, Jaeger, Zipkin, OpenCensus, and SkyWalking, Simple Log Service automatically maps the trace data to a data format defined in OpenTelemetry. For other trace data, you can use the data transformation feature to map the data to a data format supported by Simple Log Service.

Raw trace data

Raw trace data is collected to a Logstore named {instance}-traces. The following table describes the fields that are defined when the data is collected.

Field	Type	Required	Description	Example
host	String	No	The hostname of the host where the resource resides. The value is extracted from host.name of the resource field.	test-host
service	String	Yes	The service name of the resource. The value is extracted from service.name of the resource field.	test-service
resource	JSON Object	No	Other resource-related fields. The host and service fields are also resource-related fields. Examples: process ID, process name, and pod name. For more information, see Resource Semantic Conventions.	{"k8s.pod.name":"xxxx", "k8s.pod.namespace":"kube-system"}
otlp.name	String	No	The name of Trace SDK.	go-sdk
otlp.version	String	No	The version of Trace SDK.	v1.0.0
name	String	Yes	The name of the span.	/get/314159
kind	String	No	The type of the span. Examples: CLIENT and SERVER. For more information, see SpanKind.	SERVER
traceID	String	Yes	The ID of the trace. The value is a hexadecimal string.	0123456789abcde0123456789abcde
spanID	String	Yes	The ID of the span. The value is a hexadecimal string.	0123456789abcde
parentSpanID	String	Yes	The ID of the parent span. The value is a hexadecimal string.	0123456789abcde
links	JSON Array	No	The other spans that are associated with the trace data. For more information, see Specifying links.	[{"TraceID" : "abc", "SpanId" : "abc", "TraceState" : "", "Attributes" : { "k" : "v" } }]
logs	JSON Array	No	The log and event information that is associated with the trace data. For more information, see Add Events.	None
traceState	String	No	The trace state, which is defined in the World Wide Web Consortium (W3C) specification. For more information, see W3C Trace Context Specification.	None
start	INT	Yes	The start time. The value is a UNIX timestamp. Unit: nanoseconds.	1686294916826000000
end	INT	No	The end time. The value is a UNIX timestamp. Unit: nanoseconds.	1686294924827000000
duration	INT	Yes	The latency. The value is the difference between the start and end fields. Unit: nanoseconds.	8001000
attribute	JSON Object	Yes	The attribute information about the span, such as the URL and status code of the HTTP request. For more information, see Attribute Naming.	{"custom":"custom","host.hostname":"myhost","my-label":"myapp-type","null-value":"","service.name":"myapp"}
statusCode	String	Yes	The status code of the HTTP request. Valid values: OK, ERROR, and UNSET. The value UNSET is equivalent to the value OK.	ERROR
statusMessage	String	No	The status message of the HTTP request.	stack overflow

Data of call relationships among dimensions

Trace data includes the data of call relationships among dimensions. The data of call relationships is generated after calculation and is stored in a Logstore named {instance}-traces-deps. The following table describes the fields for the data.

Field	Type	Description
version	String	The call relationship among different dimensions. Valid values: service: the call relationship among services. service_name: the call relationship among services and methods. service_name_host: the call relationship among services, methods, and hosts. service_name_host_resource: the call relationship among services, methods, hosts, and resources.
child_host	String	The host information of the callee. This field is available only if version is service_name_host or service_name_host_resource.
child_name	String	The method of the callee. This field is available only if version is service_name, service_name_host, or service_name_host_resource.
child_resource	JSON Object	The resource information of the callee. This field is available only if version is service_name_host_resource.
child_service	String	The service name of the callee.
child_type	JSON Object	The supplementary information of the callee.
inner_percentile	String	The percentile. Simple Log Service uses the inner_percentile function to parse data to obtain the percentile.
max_latency	Double	The maximum latency for calling the method of the service.
min_latency	Double	The minimum latency for calling the method of the service.
n_status_fail	Double	The number of times that the method of the service fails to be called.
n_status_succ	Double	The number of times that the method of the service is successfully called.
parent_host	JSON Array	The host information of the caller. This field is available only if version is service_name_host or service_name_host_resource.
parent_name	JSON Array	The method of the caller. This field is available only if version is service_name, service_name_host, or service_name_host_resource.
parent_resource	JSON Object	The resource information of the caller. This field is available only if version is service_name_host_resource.
parent_service	INT	The service name of the caller. This field is available only if version is service, service_name, or service_name_host_resource.
parent_type	INT	The supplementary information of the caller.
sum_latency	INT	The cumulative latency for calling the method of the service.

Intermediate result data of aggregated metrics

Trace data includes the intermediate result data of aggregated metrics. The intermediate result data of aggregated metrics is generated after calculation and is stored in a Logstore named {instance}-traces-metrics. The following table describes the fields for the data.

Field	Type	Description
host	STRING	The host value of the span.
inner_percentile	STRING	The percentile. Simple Log Service uses the inner_percentile function to parse data to obtain the percentile.
max_latency	Double	The maximum latency for calling the method of the service.
min_latency	Double	The minimum latency for calling the method of the service.
n_status_fail	INT	The number of times that the method of the service fails to be called.
name	STRING	The name of the span.
resource	JSON Object	The resource information of the span.
service	STRING	The service name of the span.
sum_latency	Double	The cumulative latency for calling the method of the service. In most cases, this field is used together with the total field to calculate the average latency.
total	INT	The total number of times that the method of the service is called.
type	JSON Object	The supplementary information of the span. In most cases, the following items are included: parent: the information about the root node. mq: the message queue (MQ) that is used in the call. If this item is empty, no MQ is used in the call. kind: the type of the span. env: the environment information of the span. The value is extracted from resource.deployment.environment. version: the version of the span. The value is extracted from resource.service.version. db: the database that is used in the call. If this item is empty, no database is used in the call.
version	String	The type of the metric. The value is fixed as metric_info.