package skywalking.v3
Get desktop application:
View/edit binary Protocol Buffers messages
Collect performance raw data from browser.
rpc collectPerfData (, )
BrowserPerf.proto:34report once per page
report one or more web vitals data for pages, could report multiple times.
message
BrowserPerf.proto:124string service = 1
string serviceVersion = 2
Service version in browser is the Instance concept in the backend.
int64 time = 3
Perf data time, set by the backend side.
string pagePath = 4
Page path in browser is the endpoint concept in the backend Page path in the browser, mostly it is URI, without parameter
int32 fmpTime = 5
First Meaningful Paint time
int32 clsTime = 6
Cumulative Layout Shift time
int32 lcpTime = 7
Largest Contentful Paint time
report one or more resource data for pages, could report multiple times.
message
BrowserPerf.proto:141string service = 1
string serviceVersion = 2
Service version in browser is the Instance concept in the backend.
int64 time = 3
Perf data time, set by the backend side.
string pagePath = 4
Page path in browser is the endpoint concept in the backend Page path in the browser, mostly it is URI, without parameter
string name = 5
Resource URL
int32 duration = 6
Resource load time
int32 size = 7
Resource size
string protocol = 9
Resource load protocol
string resourceType = 8
Resource type
report one or more web interactions data for pages, could report multiple times.
message
BrowserPerf.proto:162string service = 1
string serviceVersion = 2
Service version in browser is the Instance concept in the backend.
int64 time = 3
Perf data time, set by the backend side.
string pagePath = 4
Page path in browser is the endpoint concept in the backend Page path in the browser, mostly it is URI, without parameter
int32 inpTime = 5
Interaction to Next Paint time
rpc collectErrorLogs (stream , )
BrowserPerf.proto:50report one or more error logs for pages, could report multiple times.
rpc collect (, )
CLRMetric.proto:32
Fetch the latest dynamic configurations of the service.
fetchConfigurations service requests the latest configuration. Expect command of Commands is: command: CDS # meaning ConfigurationDiscoveryService's response args: Include string key and string value pair. The key depends on the agent implementation. The value is the latest value in String value. The watcher of key owner takes the responsibility to convert it to the correct type or format. One reserved key is `UUID`. The value would help reducing the traffic load between agent and OAP if there is no change. Commands could be empty if no change detected based on ConfigurationSyncRequest.
string service = 1
Current agent service name.
string uuid = 2
[Optional] UUID is the unique id to represent the signature of the latest configuration. The agent usually uses its bootstrap configuration running in default, and this id could be empty when doing the 1st round sync. The agent could/is recommended to cache the UUID from the last command, and put it as the next round UUID parameter, then the OAP only returns the empty Commands if no configuration change happenned.
rpc queryPolicies (, )
Continuous.proto:31Query continuous profiling policy
message
Continuous.proto:40repeated policies = 1
current agent contains service and policies
Report the profiling task when the policy threshold is reached Use the returned task ID to perform the profiling task through EBPFProfilingService#collectProfilingData.
message
Continuous.proto:53string layer = 1
over threshold process entity
string serviceName = 2
string instanceName = 3
string processName = 4
repeated causes = 5
reached thresholds causes
int32 duration = 6
The execution duration for this triggered profiling. This is set at the agent side.
oneof targetTask
target profiling task
onCPU = 7
offCPU = 8
network = 9
rpc collect (stream , )
accesslog.proto:30message
accesslog.proto:34optional node = 1
current node information, only not null when first message or have update
optional connection = 2
local process and remote process connection information
repeated kernelLogs = 3
kernel level metrics
optional protocolLog = 4
application protocol log if the protocol is detected, the kernel logs is works the related logs otherwise, the kernel log is not related and is sent periodically
message
accesslog.proto:388(message has no fields)
Define the detected processes and report them.
rpc reportProcesses (, )
Process.proto:33Report discovered process in Rover
message
Process.proto:41repeated processes = 1
string ebpfAgentID = 2
An ID generated by eBPF agent, should be unique globally.
message
Process.proto:90repeated processes = 1
rpc keepAlive (, )
Process.proto:37Keep the process alive in the backend.
message
Process.proto:115repeated processes = 1
string ebpfAgentID = 2
An ID generated by eBPF agent, should be unique globally.
Define the Rover Process profiling task and upload profiling data.
rpc queryTasks (, )
Profile.proto:32Query profiling (start or stop) tasks
message
Profile.proto:40string roverInstanceId = 1
rover instance id
int64 latestUpdateTime = 2
latest task update time
rpc collectProfilingData (stream , )
Profile.proto:36collect profiling data
message
Profile.proto:48optional task = 1
task metadata
oneof profiling
profiling data
onCPU = 2
offCPU = 3
service EventService
Event.proto:30rpc collect (stream , )
Event.proto:34When reporting an event, you typically call the collect function twice, one for starting of the event and the other one for ending of the event, with the same UUID. There are also cases where you have both start time and end time already, for example, when exporting events from a 3rd-party system, the start time and end time are already known so that you can call the collect function only once.
message
Event.proto:38string uuid = 1
Unique ID of the event. Because an event may span a long period of time, the UUID is necessary to associate the start time with the end time of the same event.
optional source = 2
The source object that the event occurs on.
string name = 3
The name of the event. For example, `Reboot`, `Upgrade` etc.
Type type = 4
The type of the event. This field is friendly for UI visualization, where events of type `Normal` are considered as normal operations, while `Error` is considered as unexpected operations, such as `Crash` events, therefore we can mark them with different colors to be easier identified.
string message = 5
The detail of the event that describes why this event happened. This should be a one-line message that briefly describes why the event is reported. Examples of an `Upgrade` event may be something like `Upgrade from ${from_version} to ${to_version}`. It's NOT encouraged to include the detailed logs of this event, such as the exception stack trace.
map<string, string> parameters = 6
The parameters in the `message` field.
int64 startTime = 7
The start time (in milliseconds) of the event, measured between the current time and midnight, January 1, 1970 UTC. This field is mandatory when an event occurs.
int64 endTime = 8
The end time (in milliseconds) of the event. , measured between the current time and midnight, January 1, 1970 UTC. This field may be empty if the event has not stopped yet, otherwise it should be a valid timestamp after `startTime`.
string layer = 9
[Required] Since 9.0.0 Name of the layer to which the event belongs.
Define the JVM metrics report service.
rpc collect (, )
JVMMetric.proto:33
service LogReportService
Logging.proto:32Report collected logs into the OAP backend
rpc collect (stream , )
Logging.proto:36Recommend to report log data in a stream mode. The service/instance/endpoint of the log could share the previous value if they are not set. Reporting the logs of same service in the batch mode could reduce the network cost.
message
Logging.proto:42Log data is collected through file scratcher of agent. Natively, Satellite provides various ways to collect logs.
int64 timestamp = 1
[Optional] The timestamp of the log, in millisecond. If not set, OAP server would use the received timestamp as log's timestamp, or relies on the OAP server analyzer.
string service = 2
[Required] **Service**. Represents a set/group of workloads which provide the same behaviours for incoming requests. The logic name represents the service. This would show as a separate node in the topology. The metrics analyzed from the spans, would be aggregated for this entity as the service level. If this is not the first element of the streaming, use the previous not-null name as the service name.
string serviceInstance = 3
[Optional] **Service Instance**. Each individual workload in the Service group is known as an instance. Like `pods` in Kubernetes, it doesn't need to be a single OS process, however, if you are using instrument agents, an instance is actually a real OS process. The logic name represents the service instance. This would show as a separate node in the instance relationship. The metrics analyzed from the spans, would be aggregated for this entity as the service instance level.
string endpoint = 4
[Optional] **Endpoint**. A path in a service for incoming requests, such as an HTTP URI path or a gRPC service class + method signature. The logic name represents the endpoint, which logs belong.
optional body = 5
[Required] The content of the log.
optional traceContext = 6
[Optional] Logs with trace context
optional tags = 7
[Optional] The available tags. OAP server could provide search/analysis capabilities based on these.
string layer = 8
[Optional] Since 9.0.0 The layer of the service and servce instance. If absent, the OAP would set `layer`=`ID: 2, NAME: general`
Define the service reporting the extra information of the instance.
Report custom properties of a service instance.
rpc keepAlive (, )
Management.proto:40Keep the instance alive in the backend analysis. Only recommend to do separate keepAlive report when no trace and metrics needs to be reported. Otherwise, it is duplicated.
service MeterReportService
Meter.proto:29rpc collect (stream , )
Meter.proto:32Meter data is reported in a certain period. The agent/SDK should report all collected metrics in this period through one stream. The whole stream is an input data set, client should onComplete the stream per report period.
rpc collectBatch (stream , )
Meter.proto:38Reporting meter data in bulk mode as MeterDataCollection. By using this, each one in the stream would be treated as a complete input for MAL engine, comparing to `collect (stream MeterData)`, which is using one stream as an input data set.
message
Meter.proto:94repeated meterData = 1
service ProfileTask
Profile.proto:30query all sniffer need to execute profile task commands
rpc collectSnapshot (stream , )
Profile.proto:37collect dumped thread snapshot
rpc reportTaskFinish (, )
Profile.proto:41report profiling task finished
Define metrics report service. The metrics format is typically representing the data model collected in the service mesh case.
rpc collect (stream , )
service-mesh.proto:32message
service-mesh.proto:36oneof metrics
httpMetrics = 1
tcpMetrics = 2
message
service-mesh.proto:133(message has no fields)
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// ebpf agent(SkyWalking Rover) collects extra information from the OS(Linux Only) level to attach on the traced span. Since v3.1 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
rpc collect (stream , )
Tracing.proto:235Collect SpanAttachedEvent to the OAP server in the streaming mode.
message
Tracing.proto:244SpanAttachedEvent represents an attached event for a traced RPC. When an RPC is being traced by the in-process language agent, a span would be reported by the client-side agent. And the rover would be aware of this RPC due to the existing tracing header. Then, the rover agent collects extra information from the OS level to provide assistance information to diagnose network performance.
optional startTime = 1
The nanosecond timestamp of the event's start time. Notice, most unit of timestamp in SkyWalking is milliseconds, but NANO-SECOND is required here. Because the attached event happens in the OS syscall level, most of them are executed rapidly.
string event = 2
The official event name. For example, the event name is a method signature from syscall stack.
optional endTime = 3
[Optional] The nanosecond timestamp of the event's end time.
repeated tags = 4
The tags for this event includes some extra OS level information, such as 1. net_device used for this exit span. 2. network L7 protocol
repeated summary = 5
The summary of statistics during this event. Each statistic provides a name(metric name) to represent the name, and an int64/long as the value.
optional traceContext = 6
Refer to a trace context decoded from `sw8` header through network, such as HTTP header, MQ metadata https://skywalking.apache.org/docs/main/next/en/protocols/skywalking-cross-process-propagation-headers-protocol-v3/#standard-header-item
Define a trace segment report service. All language agents or any trace collecting component, could use this service to send span collection to the SkyWalking OAP backend.
rpc collect (stream , )
Tracing.proto:37Recommended trace segment report channel. gRPC streaming provides better performance. All language agents should choose this.
rpc collectInSync (, )
Tracing.proto:44An alternative for trace report by using gRPC unary This is provided for some 3rd-party integration, if and only if they prefer the unary mode somehow. The performance of SkyWalking OAP server would be very similar with streaming report, the performance of the network and client side are affected
message
accesslog.proto:72Connection information
Used in:
optional local = 1
local address
optional remote = 2
remote/peer address
DetectPoint role = 3
local address detect point
AccessLogConnectionTLSMode tlsMode = 4
is the connection using TLS or not
AccessLogProtocolType protocol = 5
application protocol type
optional attachment = 6
the attachment information about the connection
Used in:
Plain = 0
TLS = 1
message
accesslog.proto:156Used in:
optional startTime = 1
first bytes receive/write timestamp
optional endTime = 2
last bytes receive/write timestamp
AccessLogHTTPProtocolVersion version = 3
optional request = 4
optional response = 5
message
accesslog.proto:171Used in:
AccessLogHTTPProtocolRequestMethod method = 1
string path = 2
The path portion from the incoming request URI.
uint64 sizeOfHeadersBytes = 3
Request header and body size.
uint64 sizeOfBodyBytes = 4
optional trace = 5
The trace information if detected.
string host = 6
The host header of the incoming request.
Used in:
Get = 0
Post = 1
Put = 2
Delete = 3
Head = 4
Patch = 5
Options = 6
Trace = 7
Connect = 8
message
accesslog.proto:184Used in:
int32 statusCode = 1
uint64 sizeOfHeadersBytes = 3
Response header and body size.
uint64 sizeOfBodyBytes = 4
Used in:
HTTP1 = 0
HTTP2 = 1
message
accesslog.proto:233Used in:
optional startTime = 1
Starting to accept socket timestamp
optional endTime = 2
Finish accept operation timestamp
message
accesslog.proto:240Used in:
optional startTime = 1
Starting to close the connection timestamp
optional endTime = 2
Finish close operation timestamp
bool success = 3
Is the close operation success or not
message
accesslog.proto:224Used in:
optional startTime = 1
Starting to connect with peer address timestamp
optional endTime = 2
Finish connect operation timestamp
bool success = 3
Is the connect operation success or not
message
accesslog.proto:140Used in:
oneof operation
connect = 1
accept = 2
close = 3
read = 4
write = 5
message
accesslog.proto:323Used in:
uint32 ifindex = 1
from network hardware interface index, get the net hardware name through EBPFAccessLogNodeNetInterface#index
uint32 totalPackageCount = 2
total package count
uint64 totalPackageSize = 3
total package size
uint64 totalPackageToQueueDuration = 4
total duration(nanosecond) of all data packets queued (waiting to be read).
uint64 totalRcvPackageFromQueueDuration = 5
total duration(nanosecond) of all data packets read from the queue.
message
accesslog.proto:336Used in:
uint64 totalDuration = 1
uint64 totalRecvDuration = 2
total local receive use duration(nanoseconds) -> kernel: ip_rcv_finish - ip_rcv
uint64 totalLocalDuration = 3
total local use duration(nanoseconds) -> kernel: ip_local_deliver_finish - ip_local_deliver
uint64 totalNetFilterCount = 4
total netfiltering count and duration(nanosecond) -> kernel: nf_hook
uint64 totalNetFilterDuration = 5
message
accesslog.proto:347Used in:
uint64 totalDuration = 1
message
accesslog.proto:309Used in:
optional startTime = 1
Starting to read data timestamp
optional endTime = 2
Finish read operation timestamp
AccessLogKernelReadSyscall syscall = 3
Which kind of syscall of current read operation
optional l2Metrics = 4
Layer 2-4 related metrics
optional l3Metrics = 5
optional l4Metrics = 6
Used in:
Read = 0
Readv = 1
Recv = 2
RecvFrom = 3
RecvMsg = 4
RecvMmsg = 5
message
accesslog.proto:296Used in:
uint64 totalDuration = 1
total duration(nanosecond) of layer 2
uint32 ifindex = 2
target network hardware interface index, get the net hardware name through EBPFAccessLogNodeNetInterface#index
uint64 totalEnterQueueBufferCount = 3
total enter the net device buffer count
uint64 totalReadySendDuration = 4
total buffer data ready to send duration(nanosecond), ready_to_send - enter_queue
uint64 totalNetDeviceSendDuration = 5
total send buffer to the net device duration(nanosecond)
message
accesslog.proto:281Used in:
uint64 totalDuration = 1
total duration(nanosecond) of layer 3
uint64 totalLocalDuration = 2
total local out use duration(nanoseconds) -> kernel: ip_local_out
uint64 totalOutputDuration = 3
total output use duration(nanoseconds) -> kernel: ip_finish_output2 - ip_output
uint64 totalResolveMACCount = 5
total resolve remote MAC address(ARP Request) count and duration(nanosecond) -> kernel: neigh_resolve_output
uint64 totalResolveMACDuration = 6
uint64 totalNetFilterCount = 7
total netfiltering count and duration(nanosecond) -> kernel: nf_hook
uint64 totalNetFilterDuration = 8
message
accesslog.proto:263Used in:
uint64 totalDuration = 1
total duration(nanosecond) of layer 4
int64 totalTransmitPackageCount = 2
total send package(contains retransmit count) count(sk_buff in linux)
int64 totalRetransmitPackageCount = 3
total retransmit package count(sk_buff in linux)
repeated lossPackageMetrics = 4
total losted package count metrics
int64 totalPackageSize = 5
total package size(bytes)
message
accesslog.proto:249Used in:
optional startTime = 1
Starting to write data timestamp
optional endTime = 2
Finish write operation timestamp
AccessLogKernelWriteSyscall syscall = 3
Which kind of syscall of current write operation
optional l4Metrics = 4
Layer 2-4 related metrics
optional l3Metrics = 5
optional l2Metrics = 6
Used in:
Write = 0
Writev = 1
Send = 2
SendTo = 3
SendMsg = 4
SendMmsg = 5
SendFile = 6
SendFile64 = 7
message
accesslog.proto:276Used in:
string location = 1
int32 count = 2
message
accesslog.proto:150Used in:
oneof protocol
http = 1
Used in:
TCP = 0
HTTP_1 = 1
HTTP_2 = 2
message
accesslog.proto:191Used in:
AccessLogTraceInfoProvider provider = 1
string traceId = 2
[Optional] A string id represents the whole trace.
string traceSegmentId = 3
A unique id represents this segment. Other segments could use this id to reference as a child segment. [Optional] when this span reference
string spanId = 4
If type == SkyWalking The number id of the span. Should be unique in the whole segment. Starting at 0 If type == Zipkin The type of span ID is string.
Used in:
Zipkin = 0
SkyWalking = 1
message
BrowserPerf.proto:93Used as request type in: BrowserPerfService.collectErrorLogs, BrowserPerfService.collectErrorLogs
string uniqueId = 1
UUID
string service = 2
string serviceVersion = 3
Service version in browser is the Instance concept in the backend.
int64 time = 4
Error log time, set by the backend side.
string pagePath = 5
Page path in browser is the endpoint concept in the backend Page path in the browser, mostly it is URI, without parameter
ErrorCategory category = 6
string grade = 7
string message = 8
int32 line = 9
int32 col = 10
string stack = 11
string errorUrl = 12
bool firstReportedError = 13
Then the PV with error is only calculated when firstReportedError is true.
message
BrowserPerf.proto:54Used as request type in: BrowserPerfService.collectPerfData, BrowserPerfService.collectPerfData
string service = 1
string serviceVersion = 2
Service version in browser is the Instance concept in the backend.
int64 time = 3
Perf data time, set by the backend side.
string pagePath = 4
Page path in browser is the endpoint concept in the backend Page path in the browser, mostly it is URI, without parameter
int32 redirectTime = 5
Unit of all time related field should be `ms`.
int32 dnsTime = 6
DNS query time
int32 ttfbTime = 7
Time to first Byte
int32 tcpTime = 8
TCP connection time
int32 transTime = 9
Content transfer time
int32 domAnalysisTime = 10
Dom parsing time
int32 fptTime = 11
First paint time or blank screen time
int32 domReadyTime = 12
Dom ready time
int32 loadPageTime = 13
Page full load time
int32 resTime = 14
Synchronous load resources in the page
int32 sslTime = 15
Only valid for HTTPS
int32 ttlTime = 16
Time to interact
int32 firstPackTime = 17
First pack time
int32 fmpTime = 18
First Meaningful Paint
message
CLRMetric.proto:42Used in:
int64 time = 1
optional cpu = 2
optional gc = 3
optional thread = 4
message
CLRMetric.proto:36Used as request type in: CLRMetricReportService.collect, CLRMetricReportService.collect
repeated metrics = 1
string service = 2
string serviceInstance = 3
message
Common.proto:40Used in: ,
double usagePercent = 2
message
JVMMetric.proto:107See: https://docs.oracle.com/javase/8/docs/api/java/lang/management/ClassLoadingMXBean.html
Used in:
int64 loadedClassCount = 1
int64 totalUnloadedClassCount = 2
int64 totalLoadedClassCount = 3
message
CLRMetric.proto:49Used in:
int64 Gen0CollectCount = 1
int64 Gen1CollectCount = 2
int64 Gen2CollectCount = 3
int64 HeapMemory = 4
message
CLRMetric.proto:56Used in:
int32 AvailableCompletionPortThreads = 1
int32 AvailableWorkerThreads = 2
int32 MaxCompletionPortThreads = 3
int32 MaxWorkerThreads = 4
message
Command.proto:99Command represents an protocol customized data when return. When the agent communicates with the OAP side using gRPC, the OAP uses Command to return the data content to the Agent. The available commands are, Name: ConfigurationDiscoveryCommand Args: SerialNumber: String UUID: String properties: Key-value pairs rely on agent-side implementations Ref, Java agent supported configurations, https://skywalking.apache.org/docs/skywalking-java/next/en/setup/service-agent/java-agent/configuration-discovery/ Name: ProfileTaskQuery Args: SerialNumber: String TaskId: String EndpointName: String Duration: Integer MinDurationThreshold: Integer DumpPeriod: Integer MaxSamplingCount: Integer StartTime: Date Timestamp CreateTime: Date Timestamp Name: EBPFProfilingTaskQuery Args: TaskId: String ProcessId: Integer List TaskUpdateTime: Date timestamp TriggerType: Enum, value = FIXED_TIME TargetType: Enum, value = ON_CPU, OFF_CPU or NETWORK TaskStartTime: Date Timestamp ExtensionConfigJSON: JSON serialization of NetworkSamplings. --- NetworkSamplings --- NetworkSamplings: List URIRegex: String MinDuration: Integer When4xx: Boolean When5xx: Boolean Settings: Object RequireCompleteRequest: Boolean MaxRequestSize: Integer RequireCompleteResponse: Boolean MaxResponseSize: Integer ------------------------ FixedTriggerDuration: Long Name: ContinuousProfilingPolicyQuery Args: ServiceWithPolicyJSON: List JSON serialization of ServiceWithPolicy. --- ServiceWithPolicy --- ServiceName: String UUID: String Profiling: Multiple profiling configuration. Map Key: Profiling type. Enum, value = ON_CPU, OFF_CPU, NETWORK Value: Profiling policies. Map Key: Monitoring type. Enum, value = PROCESS_CPU, PROCESS_THREAD_COUNT, SYSTEM_LOAD, HTTP_ERROR_RATE, HTTP_AVG_RESPONSE_TIME Value: Policy configuration. Object. Threshold: String Period(s): Integer Count: Integer URIList: List<String> URIRegex: String --------------------------- Name: ContinuousProfilingReportTask Args: TaskId: String
Used in:
string command = 1
Use command name to distinguish different data type.
repeated args = 2
Data content in command. The value of content needs to be serialized as string for transmission. Basic data type: convert as string. The list of basic data: multiple data are split by ",". Complex data: serialize string through json.
message
Command.proto:112Transferring multiple Command in agent and OAP.
Used as response type in: BrowserPerfService.collectErrorLogs, BrowserPerfService.collectPerfData, CLRMetricReportService.collect, JVMMetricReportService.collect, ManagementService.keepAlive, ManagementService.reportInstanceProperties, MeterReportService.collect, ProfileTask.collectSnapshot, ProfileTask.getProfileTaskCommands, ProfileTask.reportTaskFinish, TraceSegmentReportService.collect, TraceSegmentReportService.collectInSync, v10.AsyncProfilerTask.getAsyncProfilerTaskCommands, BrowserPerfService.collectErrorLogs, BrowserPerfService.collectPerfData, BrowserPerfService.collectResourcePerfData, BrowserPerfService.collectWebInteractionsPerfData, BrowserPerfService.collectWebVitalsPerfData, CLRMetricReportService.collect, ConfigurationDiscoveryService.fetchConfigurations, ContinuousProfilingService.queryPolicies, ContinuousProfilingService.reportProfilingTask, EBPFProcessService.keepAlive, EBPFProfilingService.collectProfilingData, EBPFProfilingService.queryTasks, EventService.collect, JVMMetricReportService.collect, LogReportService.collect, ManagementService.keepAlive, ManagementService.reportInstanceProperties, MeterReportService.collect, MeterReportService.collectBatch, ProfileTask.collectSnapshot, ProfileTask.getProfileTaskCommands, ProfileTask.reportTaskFinish, SpanAttachedEventReportService.collect, TraceSegmentReportService.collect, TraceSegmentReportService.collectInSync
repeated commands = 1
message
accesslog.proto:113Used in:
oneof address
kubernetes = 1
if the address is monitored under the local machine, then return the kubernetes
ip = 2
if the address cannot be aware, then return the ip address
message
accesslog.proto:87Used in:
oneof environment
detect contains the environment information
zTunnel = 1
detect the connection have a zTunnel environment(ambient istio)
message
Continuous.proto:121eBPF Network profiling task
Used in:
repeated string samplingURIRegexes = 1
message
Continuous.proto:117eBPF off CPU profiling task
Used in:
(message has no fields)
message
Continuous.proto:113eBPF on CPU profiling task
Used in:
(message has no fields)
message
Continuous.proto:74Used in:
ContinuousProfilingTriggeredMonitorType type = 1
oneof cause
singleValue = 2
uri = 3
message
Continuous.proto:45Used in:
string serviceName = 1
service name of the each process
string uuid = 2
UUID represents the version(hash/shasum) of the current policies for the service. This is blank in the initialization stage and is set through the `ContinuousProfilingPolicyQuery` command response for the following rounds of queries.
message
Continuous.proto:98Used in:
double threshold = 1
double current = 2
Used in:
ProcessCPU = 0
Current process CPU usage percent(0-100).
ProcessThreadCount = 1
The number of threads in the process.
SystemLoad = 2
System load value.
HTTPErrorRate = 3
Process response error rate(0-100). HTTP response codes in [500-600) are considered errors. Formula: ErrorCount / TotalCount * 100
HTTPAvgResponseTime = 4
Process response average time(ms). Formula: TotalResponseTime(ms) / TotalCount
message
Continuous.proto:103Used in:
oneof uri
string regex = 1
string path = 2
double threshold = 3
double current = 4
In most cases, detect point should be `server` or `client`. Even in service mesh, this means `server`/`client` side sidecar `proxy` is reserved only.
Used in: , ,
client = 0
server = 1
proxy = 2
message
accesslog.proto:47Used in:
string name = 1
Node name
repeated netInterfaces = 2
All net interfaces list
optional bootTime = 3
System boot time
string clusterName = 4
Cluster name
optional policy = 5
Policy for the access log
message
accesslog.proto:65Used in:
int32 index = 1
int32 mtu = 2
string name = 3
message
accesslog.proto:60Used in:
repeated string excludeNamespaces = 1
Which namespaces should be excluded to generate the connection
message
Process.proto:104Used in:
int32 pid = 1
optional entityMetadata = 2
message
Process.proto:55Used in:
optional entity = 1
[required] Entity metadata Must ensure that entity information is unique at the time of reporting
int32 pid = 2
[required] The Process id of the host
repeated properties = 3
[optional] properties of the process
message
Process.proto:110Kubernetes process downstream
Used in:
int32 pid = 1
optional entityMetadata = 2
message
Process.proto:80Kubernetes process metadata
Used in:
optional entity = 1
[required] Entity metadata Must ensure that entity information is unique at the time of reporting
int32 pid = 2
[required] The Process id of the host
repeated properties = 3
[optional] properties of the process
message
Profile.proto:90Used in:
repeated stacks = 1
stack data in one task(thread)
int32 switchCount = 2
total count of the process is switched to off cpu by the scheduler.
int64 duration = 3
where time(nanoseconds) is spent waiting while blocked on I/O, locks, timers, paging/swapping, etc.
message
accesslog.proto:384Get nanosecond through offset with node start instant
Used in:
uint64 offset = 1
message
Profile.proto:83Used in:
repeated stacks = 1
stack data in one task(thread)
int32 dumpCount = 2
stack counts
message
Process.proto:94Used in:
string processId = 1
Generated process id
oneof process
Locate the process by basic information
hostProcess = 2
k8sProcess = 3
message
Process.proto:66Process Entity metadata
Used in: , , , ,
string layer = 1
[required] Process belong layer name which define in the backend
string serviceName = 2
[required] Process belong service name
string instanceName = 3
[required] Process belong service instance name
string processName = 4
[required] Process name
repeated string labels = 5
Process labels for aggregate from service
message
Process.proto:121Used in:
optional entityMetadata = 1
Process entity
repeated properties = 2
Minimize necessary properties
message
Process.proto:47Used in:
oneof metadata
The Process metadata
hostProcess = 1
k8sProcess = 2
message
Profile.proto:69Used in: ,
EBPFProfilingStackType stackType = 1
stack type
int32 stackId = 2
stack id from kernel provide
repeated string stackSymbols = 3
stack symbols
Used in:
PROCESS_KERNEL_SPACE = 0
PROCESS_USER_SPACE = 1
message
Profile.proto:58Used in:
string taskId = 1
profiling task id
string processId = 2
profiling process id
int64 profilingStartTime = 3
the start time of this profiling process
int64 currentTime = 4
report time
message
accesslog.proto:377Used in: , , , , ,
oneof timestamp
offset = 1
Used in:
ajax = 0
resource = 1
vue = 2
promise = 3
js = 4
unknown = 5
message
JVMMetric.proto:83Used in:
GCPhase phase = 1
int64 count = 2
int64 time = 3
Used in:
NEW = 0
OLD = 1
NORMAL = 2
The type of GC doesn't have new and old phases, like Z Garbage Collector (ZGC)
message
service-mesh.proto:43Used in:
int64 startTime = 1
Start timestamp in milliseconds of this RPC, measured between the current time and midnight, January 1, 1970 UTC.
int64 endTime = 2
End timestamp in milliseconds of this RPC, measured between the current time and midnight, January 1, 1970 UTC.
string sourceServiceName = 3
string sourceServiceInstance = 4
string destServiceName = 5
string destServiceInstance = 6
string endpoint = 7
int32 latency = 8
int32 responseCode = 9
bool status = 10
Status represents the response status of this calling.
Protocol protocol = 11
DetectPoint detectPoint = 12
string tlsMode = 13
NONE, mTLS, or TLS
string internalErrorCode = 14
The sidecar/proxy internal error code, the value bases on the implementation. The envoy internal error codes are listed here, https://www.envoyproxy.io/docs/envoy/latest/api-v2/data/accesslog/v2/accesslog.proto#data-accesslog-v2-responseflags
int64 internalRequestLatencyNanos = 15
Internal request(response) latency is the time cost inside the Envoy sidecar proxy. For more details, check https://www.envoyproxy.io/docs/envoy/v1.22.0/api-v3/data/accesslog/v3/accesslog.proto.html#data-accesslog-v3-accesslogcommon. internalRequestLatencyNanos is the interval between the first downstream byte received and the first upstream byte sent
int64 internalResponseLatencyNanos = 16
internalResponseLatencyNanos = time_to_first_downstream_tx_byte - time_to_first_upstream_rx_byte; | ... request workflow ... | upstream | ----- response -----> | proxy | ------- response -------> | client | | time_to_first_upstream_rx_byte | | time_to_last_downstream_tx_byte | | internalResponseLatencyNanos |
repeated sourceInstanceProperties = 17
repeated destInstanceProperties = 18
message
service-mesh.proto:82Used in:
repeated metrics = 1
message
Tracing.proto:193A ID could be represented by multiple string sections.
repeated string id = 1
message
accesslog.proto:130Used in:
string host = 1
int32 port = 2
message
Management.proto:53Used as request type in: ManagementService.keepAlive, ManagementService.keepAlive
string service = 1
string serviceInstance = 2
string layer = 3
Instance belong layer name which define in the backend, general is default.
message
Management.proto:45Used as request type in: ManagementService.reportInstanceProperties, ManagementService.reportInstanceProperties
string service = 1
string serviceInstance = 2
repeated properties = 3
string layer = 4
Instance belong layer name which define in the backend, general is default.
message
Common.proto:58since v3.1 An instantaneous point on the time-line. An instant represents a data point accurate to the nanosecond. It is constituted by a long representing epoch-seconds and an int representing nanosecond-of-second, which will always be between 0 and 999,999,999
Used in: ,
int64 seconds = 1
The number of seconds from the epoch of 1970-01-01T00:00:00Z.
int32 nanos = 2
The number of nanoseconds, later along the time-line, from the seconds field. This is always positive, and never exceeds 999,999,999.
message
Logging.proto:93JSON formatted log. The json field represents the string that could be formatted as a JSON object.
Used in:
string json = 1
message
JVMMetric.proto:43Used in:
int64 time = 1
optional cpu = 2
repeated memory = 3
repeated memoryPool = 4
repeated gc = 5
optional thread = 6
optional clazz = 7
message
JVMMetric.proto:37Used as request type in: JVMMetricReportService.collect, JVMMetricReportService.collect
repeated metrics = 1
string service = 2
string serviceInstance = 3
message
Common.proto:35since v3.1 An extension of KeyStringValuePair represents a kind of metric value for the key.
Used in:
string key = 1
int64 value = 2
message
Common.proto:28Used in: , , , , , , , , , ,
string key = 1
string value = 2
message
accesslog.proto:122Used in:
string serviceName = 1
string podName = 2
string containerName = 3
string processName = 4
int32 port = 5
message
Meter.proto:43Label of the meter
Used in: ,
string name = 1
string value = 2
message
Tracing.proto:174Used in:
int64 time = 1
The timestamp in milliseconds of this event., measured between the current time and midnight, January 1, 1970 UTC.
repeated data = 2
String key, String value pair.
message
Logging.proto:75The content of the log data
Used in:
string type = 1
A type to match analyzer(s) at the OAP server. The data could be analyzed at the client side, but could be partial
oneof content
Content with extendable format.
text = 2
json = 3
yaml = 4
message
Logging.proto:113Used in:
repeated data = 1
String key, String value pair.
message
JVMMetric.proto:53Used in:
bool isHeap = 1
int64 init = 2
int64 max = 3
int64 used = 4
int64 committed = 5
message
JVMMetric.proto:61Used in:
PoolType type = 1
int64 init = 2
int64 max = 3
int64 used = 4
int64 committed = 5
message
Meter.proto:49The histogram element definition. It includes the bucket lower boundary and the count in the bucket.
Used in:
double bucket = 1
The value represents the min value of the bucket, the upper boundary is determined by next MeterBucketValue$bucket, if it doesn't exist, the upper boundary is positive infinity.
int64 count = 2
bool isNegativeInfinity = 3
If is negative infinity, the value of the bucket is invalid
message
Meter.proto:80Single meter data, if the same metrics have a different label, they will separate.
Used as request type in: MeterReportService.collect, MeterReportService.collect
Used as field type in:
oneof metric
Meter data could be a single value or histogram.
singleValue = 1
histogram = 2
string service = 3
Service name, be set value in the first element in the stream-call.
string serviceInstance = 4
Service instance name, be set value in the first element in the stream-call.
int64 timestamp = 5
Meter data report time, be set value in the first element in the stream-call.
message
Meter.proto:70Histogram
Used in:
string name = 1
Meter name
repeated labels = 2
Labels
repeated values = 3
Customize the buckets
message
Meter.proto:60Meter single value
Used in:
string name = 1
Meter name
repeated labels = 2
Labels
double value = 3
Single value
Used in:
CODE_CACHE_USAGE = 0
NEWGEN_USAGE = 1
OLDGEN_USAGE = 2
SURVIVOR_USAGE = 3
PERMGEN_USAGE = 4
METASPACE_USAGE = 5
ZHEAP_USAGE = 6
COMPRESSED_CLASS_SPACE_USAGE = 7
CODEHEAP_NON_NMETHODS_USAGE = 8
CODEHEAP_PROFILED_NMETHODS_USAGE = 9
CODEHEAP_NON_PROFILED_NMETHODS_USAGE = 10
message
Profile.proto:46Used as request type in: ProfileTask.getProfileTaskCommands, ProfileTask.getProfileTaskCommands
string service = 1
current sniffer information
string serviceInstance = 2
int64 lastCommandTime = 3
last command timestamp
message
Profile.proto:75profile task finished report
Used as request type in: ProfileTask.reportTaskFinish, ProfileTask.reportTaskFinish
string service = 1
current sniffer information
string serviceInstance = 2
string taskId = 3
profile task
Used in:
HTTP = 0
gRPC = 1
Type of the reference
Used in:
CrossProcess = 0
Map to the reference targeting the segment in another OS process.
CrossThread = 1
Map to the reference targeting the segment in the same process of the current one, just across thread. This is only used when the coding language has the thread concept.
message
Tracing.proto:225The segment collections for trace report in batch and sync mode.
Used as request type in: TraceSegmentReportService.collectInSync, TraceSegmentReportService.collectInSync
repeated segments = 1
message
Tracing.proto:54The segment is a collection of spans. It includes all collected spans in a simple one request context, such as a HTTP request process. We recommend the agent/SDK report all tracked data of one request once for all. Typically, such as in Java, one segment represent all tracked operations(spans) of one request context in the same thread. At the same time, in some language there is not a clear `thread` concept like golang. Then, it could represent all tracked operations of one request context cross threads/goroutines.
Used as request type in: TraceSegmentReportService.collect, TraceSegmentReportService.collect
Used as field type in:
string traceId = 1
A string id represents the whole trace.
string traceSegmentId = 2
A unique id represents this segment. Other segments could use this id to reference as a child segment.
repeated spans = 3
Span collections included in this segment.
string service = 4
**Service**. Represents a set/group of workloads which provide the same behaviours for incoming requests. The logic name represents the service. This would show as a separate node in the topology. The metrics analyzed from the spans, would be aggregated for this entity as the service level.
string serviceInstance = 5
**Service Instance**. Each individual workload in the Service group is known as an instance. Like `pods` in Kubernetes, it doesn't need to be a single OS process, however, if you are using instrument agents, an instance is actually a real OS process. The logic name represents the service instance. This would show as a separate node in the instance relationship. The metrics analyzed from the spans, would be aggregated for this entity as the service instance level.
bool isSizeLimited = 6
Whether the segment includes all tracked spans. In the production environment tracked, some tasks could include too many spans for one request context, such as a batch update for a cache, or an async job. The agent/SDK could optimize or ignore some tracked spans for better performance. In this case, the value should be flagged as TRUE.
message
Tracing.proto:80Segment reference represents the link between two existing segment.
Used in:
RefType refType = 1
Represent the reference type. It could be across thread or across process. Across process means there is a downstream RPC call for this. Typically, refType == CrossProcess means SpanObject#spanType = entry.
string traceId = 2
A string id represents the whole trace.
string parentTraceSegmentId = 3
Another segment id as the parent.
int32 parentSpanId = 4
The span id in the parent trace segment.
string parentService = 5
The service logic name of the parent segment. If refType == CrossThread, this name is as same as the trace segment.
string parentServiceInstance = 6
The service logic name instance of the parent segment. If refType == CrossThread, this name is as same as the trace segment.
string parentEndpoint = 7
The endpoint name of the parent segment. **Endpoint**. A path in a service for incoming requests, such as an HTTP URI path or a gRPC service class + method signature. In a trace segment, the endpoint name is the name of first entry span.
string networkAddressUsedAtPeer = 8
The network address, including ip/hostname and port, which is used in the client side. Such as Client --> use 127.0.11.8:913 -> Server then, in the reference of entry span reported by Server, the value of this field is 127.0.11.8:913. This plays the important role in the SkyWalking STAM(Streaming Topology Analysis Method) For more details, read https://wu-sheng.github.io/STAM/
message
Event.proto:82If the event occurs on a service ONLY, the `service` field is mandatory, the serviceInstance field and endpoint field are optional; If the event occurs on a service instance, the `service` and `serviceInstance` are mandatory and endpoint is optional; If the event occurs on an endpoint, `service` and `endpoint` are mandatory, `serviceInstance` is optional;
Used in:
string service = 1
string serviceInstance = 2
string endpoint = 3
message
Tracing.proto:266Used in:
SpanReferenceType type = 1
string traceId = 2
[Optional] A string id represents the whole trace.
string traceSegmentId = 3
A unique id represents this segment. Other segments could use this id to reference as a child segment. [Optional] when this span reference
string spanId = 4
If type == SKYWALKING The number id of the span. Should be unique in the whole segment. Starting at 0 If type == ZIPKIN The type of span ID is string.
Used in:
SKYWALKING = 0
ZIPKIN = 1
Map to the layer of span
Used in:
Unknown = 0
Unknown layer. Could be anything.
Database = 1
A database layer, used in tracing the database client component.
RPCFramework = 2
A RPC layer, used in both client and server sides of RPC component.
Http = 3
HTTP is a more specific RPCFramework.
MQ = 4
A MQ layer, used in both producer and consuer sides of the MQ component.
Cache = 5
A cache layer, used in tracing the cache client component.
FAAS = 6
A FAAS layer, used in function-as-a-Service platform.
message
Tracing.proto:113Span represents a execution unit in the system, with duration and many other attributes. Span could be a method, a RPC, MQ message produce or consume. In the practice, the span should be added when it is really necessary, to avoid payload overhead. We recommend to creating spans in across process(client/server of RPC/MQ) and across thread cases only.
Used in:
int32 spanId = 1
The number id of the span. Should be unique in the whole segment. Starting at 0.
int32 parentSpanId = 2
The number id of the parent span in the whole segment. -1 represents no parent span. Also, be known as the root/first span of the segment.
int64 startTime = 3
Start timestamp in milliseconds of this span, measured between the current time and midnight, January 1, 1970 UTC.
int64 endTime = 4
End timestamp in milliseconds of this span, measured between the current time and midnight, January 1, 1970 UTC.
repeated refs = 5
<Optional> In the across thread and across process, these references targeting the parent segments. The references usually have only one element, but in batch consumer case, such as in MQ or async batch process, it could be multiple.
string operationName = 6
A logic name represents this span. We don't recommend to include the parameter, such as HTTP request parameters, as a part of the operation, especially this is the name of the entry span. All statistic for the endpoints are aggregated base on this name. Those parameters should be added in the tags if necessary. If in some cases, it have to be a part of the operation name, users should use the Group Parameterized Endpoints capability at the backend to get the meaningful metrics. Read https://github.com/apache/skywalking/blob/master/docs/en/setup/backend/endpoint-grouping-rules.md
string peer = 7
Remote address of the peer in RPC/MQ case. This is required when spanType = Exit, as it is a part of the SkyWalking STAM(Streaming Topology Analysis Method). For more details, read https://wu-sheng.github.io/STAM/
SpanType spanType = 8
Span type represents the role in the RPC context.
SpanLayer spanLayer = 9
Span layer represent the component tech stack, related to the network tech.
int32 componentId = 10
Component id is a predefinited number id in the SkyWalking. It represents the framework, tech stack used by this tracked span, such as Spring. All IDs are defined in the https://github.com/apache/skywalking/blob/master/oap-server/server-bootstrap/src/main/resources/component-libraries.yml Send a pull request if you want to add languages, components or mapping defintions, all public components could be accepted. Follow this doc for more details, https://github.com/apache/skywalking/blob/master/docs/en/guides/Component-library-settings.md
bool isError = 11
The status of the span. False means the tracked execution ends in the unexpected status. This affects the successful rate statistic in the backend. Exception or error code happened in the tracked process doesn't mean isError == true, the implementations of agent plugin and tracing SDK make the final decision.
repeated tags = 12
String key, String value pair. Tags provides more informance, includes parameters. In the OAP backend analysis, some special tag or tag combination could provide other advanced features. https://github.com/apache/skywalking/blob/master/docs/en/guides/Java-Plugin-Development-Guide.md#special-span-tags
repeated logs = 13
String key, String value pair with an accurate timestamp. Logging some events happening in the context of the span duration.
bool skipAnalysis = 14
Force the backend don't do analysis, if the value is TRUE. The backend has its own configurations to follow or override this. Use this mostly because the agent/SDK could know more context of the service role.
Map to the type of span
Used in:
Entry = 0
Server side of RPC. Consumer side of MQ.
Exit = 1
Client side of RPC. Producer side of MQ.
Local = 2
A common local code execution.
message
service-mesh.proto:86Used in:
int64 startTime = 1
Start timestamp in milliseconds of this RPC, measured between the current time and midnight, January 1, 1970 UTC.
int64 endTime = 2
End timestamp in milliseconds of this RPC, measured between the current time and midnight, January 1, 1970 UTC.
string sourceServiceName = 3
string sourceServiceInstance = 4
string destServiceName = 5
string destServiceInstance = 6
DetectPoint detectPoint = 7
string tlsMode = 8
NONE, mTLS, or TLS
string internalErrorCode = 9
The sidecar/proxy internal error code, the value bases on the implementation. The envoy internal error codes are listed here, https://www.envoyproxy.io/docs/envoy/latest/api-v2/data/accesslog/v2/accesslog.proto#data-accesslog-v2-responseflags
int64 internalRequestLatencyNanos = 10
Internal request(response) latency is the time cost inside the Envoy sidecar proxy. For more details, check https://www.envoyproxy.io/docs/envoy/v1.22.0/api-v3/data/accesslog/v3/accesslog.proto.html#data-accesslog-v3-accesslogcommon. internalRequestLatencyNanos is the interval between the first downstream byte received and the first upstream byte sent
int64 internalResponseLatencyNanos = 11
internalResponseLatencyNanos = time_to_first_downstream_tx_byte - time_to_first_upstream_rx_byte; | ... request workflow ... | upstream | ----- response -----> | proxy | ------- response -------> | client | | time_to_first_upstream_rx_byte | | time_to_last_downstream_tx_byte | | internalResponseLatencyNanos |
repeated sourceInstanceProperties = 12
repeated destInstanceProperties = 13
int64 receivedBytes = 14
The received data in bytes
int64 sentBytes = 15
The sent data in bytes
message
service-mesh.proto:124Used in:
repeated metrics = 1
message
Logging.proto:88Literal text log, typically requires regex or split mechanism to filter meaningful info.
Used in:
string text = 1
message
JVMMetric.proto:96See: https://docs.oracle.com/javase/8/docs/api/java/lang/management/ThreadMXBean.html
Used in:
int64 liveCount = 1
int64 daemonCount = 2
int64 peakCount = 3
int64 runnableStateThreadCount = 4
int64 blockedStateThreadCount = 5
int64 waitingStateThreadCount = 6
int64 timedWaitingStateThreadCount = 7
message
Profile.proto:56dumped thread snapshot
Used as request type in: ProfileTask.collectSnapshot, ProfileTask.collectSnapshot
string taskId = 1
profile task id
string traceSegmentId = 2
dumped segment id
int64 time = 3
dump timestamp
int32 sequence = 4
snapshot dump sequence, start with zero
optional stack = 5
snapshot stack
message
Profile.proto:69Used in:
repeated string codeSignatures = 1
stack code signature list
message
Logging.proto:103Logs with trace context, represent agent system has injects context(IDs) into log text.
Used in:
string traceId = 1
[Optional] A string id represents the whole trace.
string traceSegmentId = 2
[Optional] A unique id represents this segment. Other segments could use this id to reference as a child segment.
int32 spanId = 3
[Optional] The number id of the span. Should be unique in the whole segment. Starting at 0.
enum Type
Event.proto:74Used in:
Normal = 0
Error = 1
message
Logging.proto:98YAML formatted log. The yaml field represents the string that could be formatted as a YAML map.
Used in:
string yaml = 1
message
accesslog.proto:95Used in:
string real_destination_ip = 1
the real destination ip address of the connection
ZTunnelAttachmentEnvironmentDetectBy by = 2
how the environment detected
ZTunnelAttachmentSecurityPolicy security_policy = 3
the security policy of the ztunnel communicate with upstream
Used in:
ZTUNNEL_OUTBOUND_FUNC = 0
Used in:
MTLS = 0
NONE = 1