Подключиться
Language / Region
© 2021 ООО «Яндекс.Облако»
Yandex Application Load Balancer

BackendGroup

Статья создана

A set of methods for managing backend groups.

JSON Representation

{
  "id": "string",
  "name": "string",
  "description": "string",
  "folderId": "string",
  "labels": "object",
  "createdAt": "string",

  //  includes only one of the fields `http`, `grpc`, `stream`
  "http": {
    "backends": [
      {
        "name": "string",
        "backendWeight": "integer",
        "loadBalancingConfig": {
          "panicThreshold": "string",
          "localityAwareRoutingPercent": "string",
          "strictLocality": true,
          "mode": "string"
        },
        "port": "string",
        "healthchecks": [
          {
            "timeout": "string",
            "interval": "string",
            "intervalJitterPercent": "number",
            "healthyThreshold": "string",
            "unhealthyThreshold": "string",
            "healthcheckPort": "string",

            // `http.backends[].healthchecks[]` includes only one of the fields `stream`, `http`, `grpc`
            "stream": {
              "send": {
                "text": "string"
              },
              "receive": {
                "text": "string"
              }
            },
            "http": {
              "host": "string",
              "path": "string",
              "useHttp2": true
            },
            "grpc": {
              "serviceName": "string"
            },
            // end of the list of possible fields`http.backends[].healthchecks[]`

          }
        ],
        "tls": {
          "sni": "string",
          "validationContext": {

            // `http.backends[].tls.validationContext` includes only one of the fields `trustedCaId`, `trustedCaBytes`
            "trustedCaId": "string",
            "trustedCaBytes": "string",
            // end of the list of possible fields`http.backends[].tls.validationContext`

          }
        },
        "useHttp2": true,

        // `http.backends[]` includes only one of the fields `targetGroups`, `storageBucket`
        "targetGroups": {
          "targetGroupIds": [
            "string"
          ]
        },
        "storageBucket": {
          "bucket": "string"
        },
        // end of the list of possible fields`http.backends[]`

      }
    ],

    // `http` includes only one of the fields `connection`, `header`, `cookie`
    "connection": {
      "sourceIp": true
    },
    "header": {
      "headerName": "string"
    },
    "cookie": {
      "name": "string",
      "ttl": "string"
    },
    // end of the list of possible fields`http`

  },
  "grpc": {
    "backends": [
      {
        "name": "string",
        "backendWeight": "integer",
        "loadBalancingConfig": {
          "panicThreshold": "string",
          "localityAwareRoutingPercent": "string",
          "strictLocality": true,
          "mode": "string"
        },
        "port": "string",
        "healthchecks": [
          {
            "timeout": "string",
            "interval": "string",
            "intervalJitterPercent": "number",
            "healthyThreshold": "string",
            "unhealthyThreshold": "string",
            "healthcheckPort": "string",

            // `grpc.backends[].healthchecks[]` includes only one of the fields `stream`, `http`, `grpc`
            "stream": {
              "send": {
                "text": "string"
              },
              "receive": {
                "text": "string"
              }
            },
            "http": {
              "host": "string",
              "path": "string",
              "useHttp2": true
            },
            "grpc": {
              "serviceName": "string"
            },
            // end of the list of possible fields`grpc.backends[].healthchecks[]`

          }
        ],
        "tls": {
          "sni": "string",
          "validationContext": {

            // `grpc.backends[].tls.validationContext` includes only one of the fields `trustedCaId`, `trustedCaBytes`
            "trustedCaId": "string",
            "trustedCaBytes": "string",
            // end of the list of possible fields`grpc.backends[].tls.validationContext`

          }
        },
        "targetGroups": {
          "targetGroupIds": [
            "string"
          ]
        }
      }
    ],

    // `grpc` includes only one of the fields `connection`, `header`, `cookie`
    "connection": {
      "sourceIp": true
    },
    "header": {
      "headerName": "string"
    },
    "cookie": {
      "name": "string",
      "ttl": "string"
    },
    // end of the list of possible fields`grpc`

  },
  "stream": {
    "backends": [
      {
        "name": "string",
        "backendWeight": "integer",
        "loadBalancingConfig": {
          "panicThreshold": "string",
          "localityAwareRoutingPercent": "string",
          "strictLocality": true,
          "mode": "string"
        },
        "port": "string",
        "healthchecks": [
          {
            "timeout": "string",
            "interval": "string",
            "intervalJitterPercent": "number",
            "healthyThreshold": "string",
            "unhealthyThreshold": "string",
            "healthcheckPort": "string",

            // `stream.backends[].healthchecks[]` includes only one of the fields `stream`, `http`, `grpc`
            "stream": {
              "send": {
                "text": "string"
              },
              "receive": {
                "text": "string"
              }
            },
            "http": {
              "host": "string",
              "path": "string",
              "useHttp2": true
            },
            "grpc": {
              "serviceName": "string"
            },
            // end of the list of possible fields`stream.backends[].healthchecks[]`

          }
        ],
        "tls": {
          "sni": "string",
          "validationContext": {

            // `stream.backends[].tls.validationContext` includes only one of the fields `trustedCaId`, `trustedCaBytes`
            "trustedCaId": "string",
            "trustedCaBytes": "string",
            // end of the list of possible fields`stream.backends[].tls.validationContext`

          }
        },
        "targetGroups": {
          "targetGroupIds": [
            "string"
          ]
        }
      }
    ],
    "connection": {
      "sourceIp": true
    }
  },
  // end of the list of possible fields

}
Field Description
id string

ID of the backend group. Generated at creation time.
name string

Name of the backend group. The name is unique within the folder. The string length in characters is 3-63.
description string

Description of the backend group. The string is 0-256 characters long.
folderId string

ID of the folder that the backend group belongs to.
labels object

Backend group labels as key:value pairs. For details about the concept, see documentation. The maximum number of labels is 64.
createdAt string (date-time)

Creation timestamp.

String in RFC3339 text format.
http object
includes only one of the fields http, grpc, stream

An HTTP backend group resource.
http.
backends[]		 object

An HTTP backend resource.
http.
backends[].
name		 string

Required. Name of the backend.

Value must match the regular expression [a-z][-a-z0-9]{1,61}[a-z0-9].
http.
backends[].
backendWeight		 integer (int64)

Backend weight. Traffic is distributed between backends of a backend group according to their weights.

Weights must be set either for all backends in a group or for none of them. Setting no weights is the same as setting equal non-zero weights for all backends.

If the weight is non-positive, traffic is not sent to the backend.
http.
backends[].
loadBalancingConfig		 object

A load balancing configuration resource.
http.
backends[].
loadBalancingConfig.
panicThreshold		 string (int64)

Threshold for panic mode.

If percentage of healthy backends in the group drops below threshold, panic mode will be activated and traffic will be routed to all backends, regardless of their health check status. This helps to avoid overloading healthy backends. For details about panic mode, see documentation.

If the value is 0, panic mode will never be activated and traffic is routed only to healthy backends at all times.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
http.
backends[].
loadBalancingConfig.
localityAwareRoutingPercent		 string (int64)

Percentage of traffic that a load balancer node sends to healthy backends in its availability zone. The rest is divided equally between other zones. For details about zone-aware routing, see documentation.

If there are no healthy backends in an availability zone, all the traffic is divided between other zones.

If strictLocality is true, the specified value is ignored. A load balancer node sends all the traffic within its availability zone, regardless of backends' health.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
http.
backends[].
loadBalancingConfig.
strictLocality		 boolean (boolean)

Specifies whether a load balancer node should only send traffic to backends in its availability zone, regardless of their health, and ignore backends in other zones.

If set to true and there are no healthy backends in the zone, the node in this zone will respond to incoming traffic with errors. For details about strict locality, see documentation.

If strict_locality is true, the value specified in localityAwareRoutingPercent is ignored.

Default value: false.
http.
backends[].
loadBalancingConfig.
mode		 string

Load balancing mode for the backend.

For detals about load balancing modes, see documentation.

A load balancing mode resource. For details about the concept, see documentation.

  • ROUND_ROBIN: Round robin load balancing mode.

All endpoints of the backend take their turns to receive requests attributed to the backend.

  • RANDOM: Random load balancing mode. Default value.

For a request attributed to the backend, an endpoint that receives it is picked at random.

  • LEAST_REQUEST: Least request load balancing mode.

To pick an endpoint that receives a request attributed to the backend, the power of two choices algorithm is used; that is, two endpoints are picked at random, and the request is sent to the one which has the fewest active requests.

  • MAGLEV_HASH: Maglev hashing load balancing mode, used only if session affinity is working for the backend group.

Each endpoint is hashed, and a hash table with 65537 rows is filled accordingly, so that every endpoint occupies the same amount of rows. An attribute of each request, specified in session affinity configuration of the backend group, is also hashed by the same function. The row with the same number as the resulting value is looked up in the table to determine the endpoint that receives the request.

If session affinity is not working for the backend group (i.e. it is not configured or the group contains more than one backend with positive weight), endpoints for backends with MAGLEV_HASH load balancing mode are picked at RANDOM instead.
http.
backends[].
port		 string (int64)

Port used by all targets to receive traffic.

Acceptable values are 0 to 65535, inclusive.
http.
backends[].
healthchecks[]		 object

A health check resource. For details about the concept, see documentation.
http.
backends[].
healthchecks[].
timeout		 string

Required. Health check timeout.

The timeout is the time allowed for the target to respond to a check. If the target doesn't respond in time, the check is considered failed.
http.
backends[].
healthchecks[].
interval		 string

Required. Base interval between consecutive health checks.
http.
backends[].
healthchecks[].
intervalJitterPercent		 number (double)
http.
backends[].
healthchecks[].
healthyThreshold		 string (int64)

Number of consecutive successful health checks required to mark an unhealthy target as healthy.

Both 0 and 1 values amount to one successful check required.

The value is ignored when a load balancer is initialized; a target is marked healthy after one successful check.

Default value: 0.
http.
backends[].
healthchecks[].
unhealthyThreshold		 string (int64)

Number of consecutive failed health checks required to mark a healthy target as unhealthy.

Both 0 and 1 values amount to one unsuccessful check required.

The value is ignored if a health check is failed due to an HTTP 503 Service Unavailable response from the target (not applicable to TCP stream health checks). The target is immediately marked unhealthy.

Default value: 0.
http.
backends[].
healthchecks[].
healthcheckPort		 string (int64)

Port used for health checks.

If not specified, the backend port (port or port) is used for health checks.

Acceptable values are 0 to 65535, inclusive.
http.
backends[].
healthchecks[].
stream		 object
http.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for TCP stream health check settings.
http.
backends[].
healthchecks[].
stream.
send		 object

Message sent to targets during TCP data transfer.

If not specified, no data is sent to the target.

A health check payload resource.
http.
backends[].
healthchecks[].
stream.
send.
text		 string

Payload text.

The string length in characters must be greater than 0.
http.
backends[].
healthchecks[].
stream.
receive		 object

Data that must be contained in the messages received from targets for a successful health check.

If not specified, no messages are expected from targets, and those that are received are not checked.

A health check payload resource.
http.
backends[].
healthchecks[].
stream.
receive.
text		 string

Payload text.

The string length in characters must be greater than 0.
http.
backends[].
healthchecks[].
http		 object
http.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for HTTP health check settings.
http.
backends[].
healthchecks[].
http.
host		 string

Value for the HTTP/1.1 Host header or the HTTP/2 :authority pseudo-header used in requests to targets.
http.
backends[].
healthchecks[].
http.
path		 string

Required. HTTP path used in requests to targets: request URI for HTTP/1.1 request line or value for the HTTP/2 :path pseudo-header.
http.
backends[].
healthchecks[].
http.
useHttp2		 boolean (boolean)

Enables HTTP/2 usage in health checks.

Default value: false, HTTP/1.1 is used.
http.
backends[].
healthchecks[].
grpc		 object
http.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for gRPC health check settings.
http.
backends[].
healthchecks[].
grpc.
serviceName		 string

Name of the gRPC service to be checked.

If not specified, overall health is checked.

For details about the concept, see GRPC Health Checking Protocol.
http.
backends[].
tls		 object

A resource for backend TLS settings.
http.
backends[].
tls.
sni		 string

Server Name Indication (SNI) string for TLS connections.
http.
backends[].
tls.
validationContext		 object

Validation context for TLS connections.

A TLS validation context resource.
http.
backends[].
tls.
validationContext.
trustedCaId		 string
http.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes
http.
backends[].
tls.
validationContext.
trustedCaBytes		 string
http.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes

X.509 certificate contents in PEM format.
http.
backends[].
useHttp2		 boolean (boolean)

Enables HTTP/2 usage in connections between load balancer nodes and backend targets.

Default value: false, HTTP/1.1 is used.
http.
backends[].
targetGroups		 object
http.backends[] includes only one of the fields targetGroups, storageBucket

A resource for target groups that belong to the backend.
http.
backends[].
targetGroups.
targetGroupIds[]		 string

Required. List of ID's of target groups that belong to the backend.

To get the ID's of all available target groups, make a list request.

Must contain at least one element.
http.
backends[].
storageBucket		 object
http.backends[] includes only one of the fields targetGroups, storageBucket

A resource for Object Storage bucket used as a backend. For details about the concept, see documentation.
http.
backends[].
storageBucket.
bucket		 string

Required. Name of the bucket.
http.
connection		 object
http includes only one of the fields connection, header, cookie

A resource for connection-based session affinity configuration.
http.
connection.
sourceIp		 boolean (boolean)

Specifies whether an IP address of the client is used to define a connection for session affinity.
http.
header		 object
http includes only one of the fields connection, header, cookie

A resource for HTTP-header-field-based session affinity configuration.
http.
header.
headerName		 string

Name of the HTTP header field that is used for session affinity.

The string length in characters must be 1-256.
http.
cookie		 object
http includes only one of the fields connection, header, cookie

A resource for cookie-based session affinity configuration.
http.
cookie.
name		 string

Name of the cookie that is used for session affinity.

The string length in characters must be 1-256.
http.
cookie.
ttl		 string

Maximum age of cookies that are generated for sessions (persistent cookies).

If not set, session cookies are used, which are stored by clients in temporary memory and are deleted on client restarts.
grpc object
includes only one of the fields http, grpc, stream

A gRPC backend group resource.
grpc.
backends[]		 object

A gRPC backend resource.
grpc.
backends[].
name		 string

Required. Name of the backend.

Value must match the regular expression [a-z][-a-z0-9]{1,61}[a-z0-9].
grpc.
backends[].
backendWeight		 integer (int64)

Backend weight. Traffic is distributed between backends of a backend group according to their weights.

Weights must be set either for all backends of a group or for none of them. Setting no weights is the same as setting equal non-zero weights for all backends.

If the weight is non-positive, traffic is not sent to the backend.
grpc.
backends[].
loadBalancingConfig		 object
Load balancing configuration for the backend.

A load balancing configuration resource.
grpc.
backends[].
loadBalancingConfig.
panicThreshold		 string (int64)

Threshold for panic mode.

If percentage of healthy backends in the group drops below threshold, panic mode will be activated and traffic will be routed to all backends, regardless of their health check status. This helps to avoid overloading healthy backends. For details about panic mode, see documentation.

If the value is 0, panic mode will never be activated and traffic is routed only to healthy backends at all times.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
grpc.
backends[].
loadBalancingConfig.
localityAwareRoutingPercent		 string (int64)

Percentage of traffic that a load balancer node sends to healthy backends in its availability zone. The rest is divided equally between other zones. For details about zone-aware routing, see documentation.

If there are no healthy backends in an availability zone, all the traffic is divided between other zones.

If strictLocality is true, the specified value is ignored. A load balancer node sends all the traffic within its availability zone, regardless of backends' health.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
grpc.
backends[].
loadBalancingConfig.
strictLocality		 boolean (boolean)

Specifies whether a load balancer node should only send traffic to backends in its availability zone, regardless of their health, and ignore backends in other zones.

If set to true and there are no healthy backends in the zone, the node in this zone will respond to incoming traffic with errors. For details about strict locality, see documentation.

If strict_locality is true, the value specified in localityAwareRoutingPercent is ignored.

Default value: false.
grpc.
backends[].
loadBalancingConfig.
mode		 string

Load balancing mode for the backend.

For detals about load balancing modes, see documentation.

A load balancing mode resource. For details about the concept, see documentation.

  • ROUND_ROBIN: Round robin load balancing mode.

All endpoints of the backend take their turns to receive requests attributed to the backend.

  • RANDOM: Random load balancing mode. Default value.

For a request attributed to the backend, an endpoint that receives it is picked at random.

  • LEAST_REQUEST: Least request load balancing mode.

To pick an endpoint that receives a request attributed to the backend, the power of two choices algorithm is used; that is, two endpoints are picked at random, and the request is sent to the one which has the fewest active requests.

  • MAGLEV_HASH: Maglev hashing load balancing mode, used only if session affinity is working for the backend group.

Each endpoint is hashed, and a hash table with 65537 rows is filled accordingly, so that every endpoint occupies the same amount of rows. An attribute of each request, specified in session affinity configuration of the backend group, is also hashed by the same function. The row with the same number as the resulting value is looked up in the table to determine the endpoint that receives the request.

If session affinity is not working for the backend group (i.e. it is not configured or the group contains more than one backend with positive weight), endpoints for backends with MAGLEV_HASH load balancing mode are picked at RANDOM instead.
grpc.
backends[].
port		 string (int64)

Port used by all targets to receive traffic.

Acceptable values are 0 to 65535, inclusive.
grpc.
backends[].
healthchecks[]		 object

A health check resource. For details about the concept, see documentation.
grpc.
backends[].
healthchecks[].
timeout		 string

Required. Health check timeout.

The timeout is the time allowed for the target to respond to a check. If the target doesn't respond in time, the check is considered failed.
grpc.
backends[].
healthchecks[].
interval		 string

Required. Base interval between consecutive health checks.
grpc.
backends[].
healthchecks[].
intervalJitterPercent		 number (double)
grpc.
backends[].
healthchecks[].
healthyThreshold		 string (int64)

Number of consecutive successful health checks required to mark an unhealthy target as healthy.

Both 0 and 1 values amount to one successful check required.

The value is ignored when a load balancer is initialized; a target is marked healthy after one successful check.

Default value: 0.
grpc.
backends[].
healthchecks[].
unhealthyThreshold		 string (int64)

Number of consecutive failed health checks required to mark a healthy target as unhealthy.

Both 0 and 1 values amount to one unsuccessful check required.

The value is ignored if a health check is failed due to an HTTP 503 Service Unavailable response from the target (not applicable to TCP stream health checks). The target is immediately marked unhealthy.

Default value: 0.
grpc.
backends[].
healthchecks[].
healthcheckPort		 string (int64)

Port used for health checks.

If not specified, the backend port (port or port) is used for health checks.

Acceptable values are 0 to 65535, inclusive.
grpc.
backends[].
healthchecks[].
stream		 object
TCP stream health check settings.
grpc.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for TCP stream health check settings.
grpc.
backends[].
healthchecks[].
stream.
send		 object

Message sent to targets during TCP data transfer.

If not specified, no data is sent to the target.

A health check payload resource.
grpc.
backends[].
healthchecks[].
stream.
send.
text		 string

Payload text.

The string length in characters must be greater than 0.
grpc.
backends[].
healthchecks[].
stream.
receive		 object

Data that must be contained in the messages received from targets for a successful health check.

If not specified, no messages are expected from targets, and those that are received are not checked.

A health check payload resource.
grpc.
backends[].
healthchecks[].
stream.
receive.
text		 string

Payload text.

The string length in characters must be greater than 0.
grpc.
backends[].
healthchecks[].
http		 object
HTTP health check settings.
grpc.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for HTTP health check settings.
grpc.
backends[].
healthchecks[].
http.
host		 string

Value for the HTTP/1.1 Host header or the HTTP/2 :authority pseudo-header used in requests to targets.
grpc.
backends[].
healthchecks[].
http.
path		 string

Required. HTTP path used in requests to targets: request URI for HTTP/1.1 request line or value for the HTTP/2 :path pseudo-header.
grpc.
backends[].
healthchecks[].
http.
useHttp2		 boolean (boolean)

Enables HTTP/2 usage in health checks.

Default value: false, HTTP/1.1 is used.
grpc.
backends[].
healthchecks[].
grpc		 object
gRPC health check settings.
grpc.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for gRPC health check settings.
grpc.
backends[].
healthchecks[].
grpc.
serviceName		 string

Name of the gRPC service to be checked.

If not specified, overall health is checked.

For details about the concept, see GRPC Health Checking Protocol.
grpc.
backends[].
tls		 object
Settings for TLS connections between load balancer nodes and backend targets. If specified, the load balancer establishes HTTPS (HTTP over TLS) connections with targets and compares received certificates with the one specified in validationContext. If not specified, the load balancer establishes unencrypted HTTP connections with targets.

A resource for backend TLS settings.
grpc.
backends[].
tls.
sni		 string

Server Name Indication (SNI) string for TLS connections.
grpc.
backends[].
tls.
validationContext		 object

Validation context for TLS connections.

A TLS validation context resource.
grpc.
backends[].
tls.
validationContext.
trustedCaId		 string
grpc.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes
grpc.
backends[].
tls.
validationContext.
trustedCaBytes		 string
grpc.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes

X.509 certificate contents in PEM format.
grpc.
backends[].
targetGroups		 object
Target groups that belong to the backend. For details about target groups, see documentation.

A resource for target groups that belong to the backend.
grpc.
backends[].
targetGroups.
targetGroupIds[]		 string

Required. List of ID's of target groups that belong to the backend.

To get the ID's of all available target groups, make a list request.

Must contain at least one element.
grpc.
connection		 object
Connection-based session affinity configuration. For now, a connection is defined only by an IP address of the client.
grpc includes only one of the fields connection, header, cookie

A resource for connection-based session affinity configuration.
grpc.
connection.
sourceIp		 boolean (boolean)

Specifies whether an IP address of the client is used to define a connection for session affinity.
grpc.
header		 object
HTTP-header-field-based session affinity configuration.
grpc includes only one of the fields connection, header, cookie

A resource for HTTP-header-field-based session affinity configuration.
grpc.
header.
headerName		 string

Name of the HTTP header field that is used for session affinity.

The string length in characters must be 1-256.
grpc.
cookie		 object
Cookie-based session affinity configuration.
grpc includes only one of the fields connection, header, cookie

A resource for cookie-based session affinity configuration.
grpc.
cookie.
name		 string

Name of the cookie that is used for session affinity.

The string length in characters must be 1-256.
grpc.
cookie.
ttl		 string

Maximum age of cookies that are generated for sessions (persistent cookies).

If not set, session cookies are used, which are stored by clients in temporary memory and are deleted on client restarts.
stream object
includes only one of the fields http, grpc, stream

A Stream backend group resource.
stream.
backends[]		 object

A stream backend resource.
stream.
backends[].
name		 string

Value must match the regular expression [a-z][-a-z0-9]{1,61}[a-z0-9].
stream.
backends[].
backendWeight		 integer (int64)

If not set, backend will be disabled.
stream.
backends[].
loadBalancingConfig		 object
Load balancing configuration for the backend.

A load balancing configuration resource.
stream.
backends[].
loadBalancingConfig.
panicThreshold		 string (int64)

Threshold for panic mode.

If percentage of healthy backends in the group drops below threshold, panic mode will be activated and traffic will be routed to all backends, regardless of their health check status. This helps to avoid overloading healthy backends. For details about panic mode, see documentation.

If the value is 0, panic mode will never be activated and traffic is routed only to healthy backends at all times.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
stream.
backends[].
loadBalancingConfig.
localityAwareRoutingPercent		 string (int64)

Percentage of traffic that a load balancer node sends to healthy backends in its availability zone. The rest is divided equally between other zones. For details about zone-aware routing, see documentation.

If there are no healthy backends in an availability zone, all the traffic is divided between other zones.

If strictLocality is true, the specified value is ignored. A load balancer node sends all the traffic within its availability zone, regardless of backends' health.

Default value: 0.

Acceptable values are 0 to 100, inclusive.
stream.
backends[].
loadBalancingConfig.
strictLocality		 boolean (boolean)

Specifies whether a load balancer node should only send traffic to backends in its availability zone, regardless of their health, and ignore backends in other zones.

If set to true and there are no healthy backends in the zone, the node in this zone will respond to incoming traffic with errors. For details about strict locality, see documentation.

If strict_locality is true, the value specified in localityAwareRoutingPercent is ignored.

Default value: false.
stream.
backends[].
loadBalancingConfig.
mode		 string

Load balancing mode for the backend.

For detals about load balancing modes, see documentation.

A load balancing mode resource. For details about the concept, see documentation.

  • ROUND_ROBIN: Round robin load balancing mode.

All endpoints of the backend take their turns to receive requests attributed to the backend.

  • RANDOM: Random load balancing mode. Default value.

For a request attributed to the backend, an endpoint that receives it is picked at random.

  • LEAST_REQUEST: Least request load balancing mode.

To pick an endpoint that receives a request attributed to the backend, the power of two choices algorithm is used; that is, two endpoints are picked at random, and the request is sent to the one which has the fewest active requests.

  • MAGLEV_HASH: Maglev hashing load balancing mode, used only if session affinity is working for the backend group.

Each endpoint is hashed, and a hash table with 65537 rows is filled accordingly, so that every endpoint occupies the same amount of rows. An attribute of each request, specified in session affinity configuration of the backend group, is also hashed by the same function. The row with the same number as the resulting value is looked up in the table to determine the endpoint that receives the request.

If session affinity is not working for the backend group (i.e. it is not configured or the group contains more than one backend with positive weight), endpoints for backends with MAGLEV_HASH load balancing mode are picked at RANDOM instead.
stream.
backends[].
port		 string (int64)

Optional alternative port for all targets.

Acceptable values are 0 to 65535, inclusive.
stream.
backends[].
healthchecks[]		 object

A health check resource. For details about the concept, see documentation.
stream.
backends[].
healthchecks[].
timeout		 string

Required. Health check timeout.

The timeout is the time allowed for the target to respond to a check. If the target doesn't respond in time, the check is considered failed.
stream.
backends[].
healthchecks[].
interval		 string

Required. Base interval between consecutive health checks.
stream.
backends[].
healthchecks[].
intervalJitterPercent		 number (double)
stream.
backends[].
healthchecks[].
healthyThreshold		 string (int64)

Number of consecutive successful health checks required to mark an unhealthy target as healthy.

Both 0 and 1 values amount to one successful check required.

The value is ignored when a load balancer is initialized; a target is marked healthy after one successful check.

Default value: 0.
stream.
backends[].
healthchecks[].
unhealthyThreshold		 string (int64)

Number of consecutive failed health checks required to mark a healthy target as unhealthy.

Both 0 and 1 values amount to one unsuccessful check required.

The value is ignored if a health check is failed due to an HTTP 503 Service Unavailable response from the target (not applicable to TCP stream health checks). The target is immediately marked unhealthy.

Default value: 0.
stream.
backends[].
healthchecks[].
healthcheckPort		 string (int64)

Port used for health checks.

If not specified, the backend port (port or port) is used for health checks.

Acceptable values are 0 to 65535, inclusive.
stream.
backends[].
healthchecks[].
stream		 object
TCP stream health check settings.
stream.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for TCP stream health check settings.
stream.
backends[].
healthchecks[].
stream.
send		 object

Message sent to targets during TCP data transfer.

If not specified, no data is sent to the target.

A health check payload resource.
stream.
backends[].
healthchecks[].
stream.
send.
text		 string

Payload text.

The string length in characters must be greater than 0.
stream.
backends[].
healthchecks[].
stream.
receive		 object

Data that must be contained in the messages received from targets for a successful health check.

If not specified, no messages are expected from targets, and those that are received are not checked.

A health check payload resource.
stream.
backends[].
healthchecks[].
stream.
receive.
text		 string

Payload text.

The string length in characters must be greater than 0.
stream.
backends[].
healthchecks[].
http		 object
HTTP health check settings.
stream.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for HTTP health check settings.
stream.
backends[].
healthchecks[].
http.
host		 string

Value for the HTTP/1.1 Host header or the HTTP/2 :authority pseudo-header used in requests to targets.
stream.
backends[].
healthchecks[].
http.
path		 string

Required. HTTP path used in requests to targets: request URI for HTTP/1.1 request line or value for the HTTP/2 :path pseudo-header.
stream.
backends[].
healthchecks[].
http.
useHttp2		 boolean (boolean)

Enables HTTP/2 usage in health checks.

Default value: false, HTTP/1.1 is used.
stream.
backends[].
healthchecks[].
grpc		 object
gRPC health check settings.
stream.backends[].healthchecks[] includes only one of the fields stream, http, grpc

A resource for gRPC health check settings.
stream.
backends[].
healthchecks[].
grpc.
serviceName		 string

Name of the gRPC service to be checked.

If not specified, overall health is checked.

For details about the concept, see GRPC Health Checking Protocol.
stream.
backends[].
tls		 object
Settings for TLS connections between load balancer nodes and backend targets. If specified, the load balancer establishes HTTPS (HTTP over TLS) connections with targets and compares received certificates with the one specified in validationContext. If not specified, the load balancer establishes unencrypted HTTP connections with targets.

A resource for backend TLS settings.
stream.
backends[].
tls.
sni		 string

Server Name Indication (SNI) string for TLS connections.
stream.
backends[].
tls.
validationContext		 object

Validation context for TLS connections.

A TLS validation context resource.
stream.
backends[].
tls.
validationContext.
trustedCaId		 string
stream.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes
stream.
backends[].
tls.
validationContext.
trustedCaBytes		 string
stream.backends[].tls.validationContext includes only one of the fields trustedCaId, trustedCaBytes

X.509 certificate contents in PEM format.
stream.
backends[].
targetGroups		 object
Target groups that belong to the backend.

A resource for target groups that belong to the backend.
stream.
backends[].
targetGroups.
targetGroupIds[]		 string

Required. List of ID's of target groups that belong to the backend.

To get the ID's of all available target groups, make a list request.

Must contain at least one element.
stream.
connection		 object
Connection-based session affinity configuration. For now, a connection is defined only by an IP address of the client.

A resource for connection-based session affinity configuration.
stream.
connection.
sourceIp		 boolean (boolean)

Specifies whether an IP address of the client is used to define a connection for session affinity.

Methods

Method Description
addBackend Adds backends to the specified backend group.
create Creates a backend group in the specified folder.
delete Deletes the specified backend group.
get Returns the specified backend group.
list Lists backend groups in the specified folder.
listOperations Lists operations for the specified backend group.
removeBackend Removes backends from the specified backend group.
update Updates the specified backend group.
updateBackend Updates the specified backend.
Language / Region
© 2021 ООО «Яндекс.Облако»
В этой статье: