5.2 KiB

Raw Blame History

Query Protocol

Query Protocol defines a set of APIs in GraphQL grammar to provide data query and interactive capabilities with SkyWalking native visualization tool or 3rd party system, including Web UI, CLI or private system.

Query protocol official repository, https://github.com/apache/skywalking-query-protocol.

Metadata

Metadata includes the brief info of the whole under monitoring services and their instances, endpoints, etc. Use multiple ways to query this meta data.

extend type Query {
    getGlobalBrief(duration: Duration!): ClusterBrief

    # Normal service related metainfo 
    getAllServices(duration: Duration!): [Service!]!
    searchServices(duration: Duration!, keyword: String!): [Service!]!
    searchService(serviceCode: String!): Service
    
    # Fetch all services of Browser type
    getAllBrowserServices(duration: Duration!): [Service!]!

    # Service intance query
    getServiceInstances(duration: Duration!, serviceId: ID!): [ServiceInstance!]!

    # Endpoint query
    # Consider there are huge numbers of endpoint,
    # must use endpoint owner's service id, keyword and limit filter to do query.
    searchEndpoint(keyword: String!, serviceId: ID!, limit: Int!): [Endpoint!]!
    getEndpointInfo(endpointId: ID!): EndpointInfo

    # Database related meta info.
    getAllDatabases(duration: Duration!): [Database!]!
    getTimeInfo: TimeInfo
}

Topology

Show the topology and dependency graph of services or endpoints. Including direct relationship or global map.

extend type Query {
    # Query the global topology
    getGlobalTopology(duration: Duration!): Topology
    # Query the topology, based on the given service
    getServiceTopology(serviceId: ID!, duration: Duration!): Topology
    # Query the instance topology, based on the given clientServiceId and serverServiceId
    getServiceInstanceTopology(clientServiceId: ID!, serverServiceId: ID!, duration: Duration!): ServiceInstanceTopology
    # Query the topology, based on the given endpoint
    getEndpointTopology(endpointId: ID!, duration: Duration!): Topology
}

Metrics

Metrics query targets all the objects defined in OAL script. You could get the metrics data in linear or thermodynamic matrix formats based on the aggregation functions in script.

3 types of metrics could be query

Single value. The type of most default metrics is single value, consider this as default. getValues and getLinearIntValues are suitable for this.
Multiple value. One metrics defined in OAL include multiple value calculations. Use getMultipleLinearIntValues to get all values. percentile is a typical multiple value func in OAL.
Heatmap value. Read Heatmap in WIKI for detail. thermodynamic is the only OAL func. Use getThermodynamic to get the values.

extend type Query {
    getValues(metric: BatchMetricConditions!, duration: Duration!): IntValues
    getLinearIntValues(metric: MetricCondition!, duration: Duration!): IntValues
    # Query the type of metrics including multiple values, and format them as multiple linears.
    # The seq of these multiple lines base on the calculation func in OAL
    # Such as, should us this to query the result of func percentile(50,75,90,95,99) in OAL,
    # then five lines will be responsed, p50 is the first element of return value.
    getMultipleLinearIntValues(metric: MetricCondition!, numOfLinear: Int!, duration: Duration!): [IntValues!]!
    getThermodynamic(metric: MetricCondition!, duration: Duration!): Thermodynamic
}

Metrics are defined in the official_analysis.oal.

Aggregation

Aggregation query means the metrics data need a secondary aggregation in query stage, which makes the query interfaces have some different arguments. Such as, TopN list of services is a very typical aggregation query, metrics stream aggregation just calculates the metrics values of each service, but the expected list needs ordering metrics data by the values.

Aggregation query is for single value metrics only.

# The aggregation query is different with the metric query.
# All aggregation queries require backend or/and storage do aggregation in query time.
extend type Query {
    # TopN is an aggregation query.
    getServiceTopN(name: String!, topN: Int!, duration: Duration!, order: Order!): [TopNEntity!]!
    getAllServiceInstanceTopN(name: String!, topN: Int!, duration: Duration!, order: Order!): [TopNEntity!]!
    getServiceInstanceTopN(serviceId: ID!, name: String!, topN: Int!, duration: Duration!, order: Order!): [TopNEntity!]!
    getAllEndpointTopN(name: String!, topN: Int!, duration: Duration!, order: Order!): [TopNEntity!]!
    getEndpointTopN(serviceId: ID!, name: String!, topN: Int!, duration: Duration!, order: Order!): [TopNEntity!]!
}

Others

The following query(s) are for specific features, including trace, alarm or profile.

Trace. Query distributed traces by this.
Alarm. Through alarm query, you can have alarm trend and details.

The actual query GraphQL scrips could be found inside query-protocol folder in here.

5.2 KiB Raw Blame History