Consul
Service resolver configuration reference
This page provides reference information for service resolver configuration entries. Configure and apply service resolvers to create named subsets of service instances and define their behavior when satisfying upstream requests.
Refer to L7 traffic management overview for additional information.
Configuration model
The following list outlines field hierarchy, language-specific data types, and requirements in the configuration entry. Click on a property name to view additional details, including default values.
Kind
: string | required | must be set toservice-resolver
Name
: string | requiredNamespace
: string |default
EnterprisePartition
: string |default
EnterpriseMeta
: mapConnectTimeout
: string |0s
RequestTimeout
: string |15s
Subsets
: mapFilter
: stringOnlyPassing
: boolean |false
DefaultSubset
: stringRedirect
: mapService
: stringServiceSubset
: stringNamespace
: string EnterprisePartition
: string |default
EnterpriseSamenessGroup
: string EnterpriseDatacenter
: stringPeer
: string
PrioritizeByLocality
: map EnterpriseMode
: string |failover
Failover
: mapService
: stringServiceSubset
: stringNamespace
: string EnterpriseSamenessGroup
: string EnterpriseDatacenters
: listTargets
: listService
: stringServiceSubset
: stringNamespace
: string EnterprisePartition
: string |default
EnterpriseDatacenter
: stringPeer
: string
LoadBalancer
: mapPolicy
: stringLeastRequestConfig
: mapChoiceCount
: integer |2
RingHashConfig
: mapMinimumRingSize
: integer |1024
MaximumRingSize
: integer |8192
HashPolicies
: mapField
: stringFieldValue
: stringCookieConfig
: mapSourceIP
: boolean |false
Terminal
: boolean |false
Complete configuration
When every field is defined, a service resolver configuration entry has the following form:
Kind = "service-resolver" ## required
Name = "<name-of-service-configuration-applies-to>"
Namespace = "<namespace-configuration-applies-to>"
Partition = "<partition-configuration-applies-to>"
Meta = {
<key> = "<value>"
}
ConnectTimeout = "10s"
RequestTimeout = "15s"
Subsets = {
<subset01-name> = {
Filter = "<expression.to.match.on == value01>"
OnlyPassing = true
}
<subset02-name> = {
Filter = "<expression.to.match.on == value02>"
OnlyPassing = true
}
}
DefaultSubset = "<subset01-name>"
Redirect = {
Service = "<destination-service>"
ServiceSubset = "<destination-subset>"
Namespace = "<destination-namespace>"
Partition = "<destination-partition>"
SamenessGroup = "<destination-sameness-group>"
Datacenter = "<destination-datacenter>"
Peer = "<destination-peer>"
}
PrioritizeByLocality = {
Mode = "failover"
}
Failover = { ## requires at least one of the following: Service, ServiceSubset, Namespace, Targets, Datacenters, SamenessGroup
<local-subset-name> = {
Targets = [
{ Service = "<destination-service>" },
{ ServiceSubset = "<destination-service-subset>" },
{ Namespace = "<destination-namespace>" },
{ Partition = "<destination-partition>" },
{ Datacenter = "<destination-datacenter>" },
{ Peer = "<destination-peer>" }
]
}
"*" = {
Service = "<destination-service>"
ServiceSubset = "<destination-service-subset>"
Namespace = "<destination-namespace>"
Datacenters = ["<destination-datacenter-01>", "<destination-datacenter-02>"]
}
}
LoadBalancer = {
Policy = "random"
LeastRequestConfig = { ## requires Policy = "least_request"
ChoiceCount = 2
}
RingHashConfig = { ## requires Policy = "ring_hash"
MinimumRingSize = 1024
MaximumRingSize = 8192
}
HashPolicies = [
{
Field = "header" ## cannot specify with SourceIP
FieldValue = "<value-to-hash>" ## cannot specify with SourceIP
CookieConfig = {
Session = false
TTL = "0s"
Path = "<path/to/cookie>"
}
SourceIP = false ## cannot specify with Field or FieldValue
Terminal = false
}
]
}
Specification
This section provides details about the fields you can configure in the configuration entry.
Kind
Specifies the type of configuration entry to implement. Must be set to service-resolver
.
Values
- Default: None
- This field is required.
- Data type: String value that must be set to
service-resolver
.
Name
Specifies a name for the configuration entry. The name is metadata that you can use to reference the configuration entry when performing Consul operations, such as applying a configuration entry to a specific cluster.
Values
- Default: None
- This field is required.
- Data type: String
Namespace
Enterprise
Specifies the namespace that the service resolver applies to. Refer to namespaces for more information.
Values
- Default: None
- Data type: String
Partition
Enterprise
Specifies the admin partition that the service resolver applies to. Refer to admin partitions for more information.
Values
- Default:
default
- Data type: String
Meta
Specifies key-value pairs to add to the KV store.
Values
ConnectTimeout
Specifies the timeout duration for establishing new network connections to this service. By default, the duration is measured in nanoseconds (ns).
Values
- Default: None
- Data type: String
RequestTimeout
Specifies the timeout duration for receiving an HTTP response from this service. When set to 0s
, the default value of 15s
is used instead. By default, the duration is measured in nanoseconds (ns).
Values
- Default:
15s
- Data type: String
Subsets
Specifies names for custom service subsets and the conditions under which service instances belong to each subset. Define a subset by specifying a key and a value where the key is the subset’s name and the value is a map that contains a filtering expression. If this parameter is not specified, only the unnamed default subset is usable.
For additional guidance, refer to the filter on service version configuration example.
Values
Default: None
Data type: Map containing a key-value pair.
<KEY>
: String that names the subset. The string must be valid as a DNS subdomain element.<VALUE>
: Map that can contain the following parameters:Parameter Description Data type Default Filter
Specifies an expression that filters the DNS elements of service instances that belong to the subset. If empty, all healthy instances of a service are returned. This expression can filter on the same DNS selectors as the Health API endpoint. For more information about creating and using expressions to filter, refer to filtering. String None OnlyPassing
Determines if instances that return a warning from a health check are allowed to resolve a request. When set to false
, instances withpassing
andwarning
states are considered healthy. When set totrue
, only instances with apassing
health check state are considered healthy.Boolean false
DefaultSubset
Specifies a defined subset of service instances to use when no explicit subset is requested. If this parameter is not specified, Consul uses the unnamed default subset.
Values
- Default: None
- Data type: String
Redirect
Specifies redirect instructions for local service traffic so that services deployed to a different network location resolve the upstream request instead. When this field is defined, Consul ignores all other fields in a service resolver configuration entry except for Kind
, Name
, Namespace
. When there are multiple redirects defined for a single service, Consul uses only the first one it applies.
Values
- Default: None
- Data type: Map that can contain the following parameters:
Redirect{}.Service
Specifies the name of a service at the redirect’s destination that resolves local upstream requests.
Values
- Default: None
- Data type: String
Redirect{}.ServiceSubset
Specifies the name of a subset of services at the redirect’s destination that resolves local upstream requests. If empty, the default subset is used. If specified, you must also specify at least one of the following in the same Redirect
map: Service
, Namespace
, andDatacenter
.
Values
- Default: None
- Data type: String
Redirect{}.Namespace
Enterprise
Specifies the namespace at the redirect’s destination that resolves local upstream requests.
Values
- Default: None
- Data type: String
Redirect{}.Partition
Enterprise
Specifies the admin partition at the redirect’s destination that resolves local upstream requests.
Values
- Default: None
- Data type: String
Redirect{}.SamenessGroup
Enterprise
Specifies the sameness group at the redirect’s destination that resolves local upstream requests.
Values
- Default: None
- Data type: String
Redirect{}.Datacenter
Specifies the datacenter at the redirect’s destination that resolves local upstream requests.
Values
- Default: None
- Data type: String
Redirect{}.Peer
Specifies the cluster with an active cluster peering connection at the redirect’s destination that resolves local upstream requests. Requires separately defined service intentions that authorize services for peers. When Consul runs a health check before resolving requests from the peer, it does not apply health checks that were defined on the peer and exported to the local cluster through the exported services configuration entry.
Values
- Default: None
- Data type: String
PrioritizeByLocality
Sets a mode for the service that allows instances to prioritize upstream targets that are in the same network region and zone. You can specify the following string values for the mode
field:
failover
: If the upstream targets that a service is connected to become unreachable, the service prioritizes healthy upstream instances with matchingLocality
configuration. Refer to Route traffic to local upstreams for additional information.
Values
- Default: None
- Data type: Map
Failover
Specifies controls for rerouting traffic to an alternate pool of service instances if the target service fails.
This parameter is a map, and its key is the name of the local service subset that resolves to another location when it fails. You can specify a "*"
wildcard to apply failovers to any subset.
Service
, ServiceSubset
, Namespace
, Targets
, SamenessGroup
, and Datacenters
cannot all be empty at the same time.
Values
- Default: None
- Data type: Map that can contain the following parameters:
Failover{}.Service
Specifies the name of the service to resolve at the failover location during a failover scenario.
Values
- Default: None
- Data type: String
Failover{}.ServiceSubset
Specifies the name of a subset of service instances to resolve at the failover location during a failover scenario. If empty, Consul uses the service’s DefaultSubset
.
Values
- Default: None
- Data type: String
Failover{}.Namespace
Enterprise
Specifies the namespace at the failover location where the failover services are deployed. If empty, the current namespace is used.
Values
- Default: None
- Data type: String
Failover{}.SamenessGroup
Enterprise
Specifies the sameness group at the failover location where the failover services are deployed.
Values
- Default: None
- Data type: String
Failover{}.Datacenters
Specifies an ordered list of datacenters at the failover location to attempt connections to during a failover scenario. When Consul cannot establish a connection with the first datacenter in the list, it proceeds sequentially until establishing a connection with another datacenter.
Values
- Default: None
- Data type: String
Failover{}.Targets
Specifies a fixed list of failover targets to try during failover. This list can express complicated failover scenarios.
For examples, refer to the failover example configurations.
Values
- Default: None
- Data type: List of maps that can contain the following parameters:
Failover{}.Targets[].Service
Specifies the service name to use for the failover target. If empty, the current service name is used.
Values
- Default: None
- Data type: String
Failover{}.Targets[].ServiceSubset
Specifies the named subset to use for the failover target. If empty, the default subset for the requested service name is used.
Values
- Default: None
- Data type: String
Failover{}.Targets[].Namespace
Enterprise
Specifies the namespace to use for the failover target. If empty, the default namespace is used.
Values
- Default: None
- Data type: String
Failover{}.Targets[].Partition
Enterprise
Specifies the admin partition within the same datacenter to use for the failover target. If empty, the default
partition is used. To use an admin partition in a different datacenter for the failover target, use the Peer
field instead.
Values
- Default: None
- Data type: String
Failover{}.Targets[].Datacenter
Specifies the WAN federated datacenter to use for the failover target. If empty, the current datacenter is used. To use a datacenter for the failover target that is connected with a cluster peering relationship rather than WAN federation, use the Peer
field instead.
Values
- Default: None
- Data type: String
Failover{}.Targets[].Peer
Specifies the destination cluster peer to resolve the target service name from. Intentions must be defined on the peered cluster so that the source service can access this failover target service as an upstream. When the peer name is specified, Consul uses Envoy's outlier detection to determine the health of the failover target based on whether services can communicate with the failover target. Consul ignores service health checks imported from a peer for failover targets because the checks do not account for service routers, splitters, and resolvers that may be defined in the peer for the target service.
Values
- Default: None
- Data type: String
LoadBalancer
Specifies the load balancing policy and configuration for services issuing requests to this upstream.
Values
- Default: None
- Data type: Map that can contain the following parameters:
LoadBalancer{}.Policy
Specifies the type of load balancing policy for selecting a host. Supported load balancing policies include:
Policy | Description |
---|---|
Random | The load balancer forwards a client request to an available server chosen at random from a fixed list. |
Round robin | The load balancer proceeds through a fixed list of servers and forwards a client request to each available server in order. |
Least request | The load balancer forwards a client request to the server with the fewest connections. When using this policy, you can specify additional parameters in LoadBalancer{}.LeastRequestConfig . |
Ring hash | The load balancer forwards requests from the same client to the same group of servers. When using this policy, you can specify additional parameters in LoadBalancer{}.RingHashConfig . |
Maglev | The load balancer uses a hashing algorithm to spread requests evenly across servers. When using this policy, you can specify additional parameters in LoadBalancer{}.HashPolicies . |
Values
Default: None
Data type: String containing one of the following values:
LoadBalancer{}.LeastRequestConfig
Specifies configuration for the least_request
policy type.
Values
Default: None
Data type: Map containing the following parameter:
Parameter Description Data type Default ChoiceCount
Specifies the number of random healthy hosts from which to select the one with the least requests. Integer 2
LoadBalancer{}.RingHashConfig
Specifies configuration for the ring_hash
policy type.
Values
Default: None
Data type: List that can contain the following parameters:
Parameter Description Data type Default MinimumRingSize
Determines the minimum number of entries in the hash ring. Integer 1024
MaximumRingSize
Determines the maximum number of entries in the hash ring. Integer 8192
LoadBalancer{}.HashPolicies
Specifies a list of hash policies to use for hashing load balancing algorithms. Consul evaluates hash policies individually and combines them so that identical lists result in the same hash. If no hash policies are present or successfully evaluated, then Consul selects a random backend host.
Values
- Default: None
- Data type: List of maps for the following parameters:
LoadBalancer{}.HashPolicies[].Field
Specifies the attribute type to hash on. You cannot specify the Field
parameter if SourceIP
is also configured.
Supported attribute types include the following:
Attribute | Description |
---|---|
Cookie | The load balancer uses a cookie to obtain a hash key. Refer to the Envoy documentation for more information. |
Header | The load balancer uses a request header to obtain a hash key. Refer to the Envoy documentation for more information. |
Query Parameter | The load balancer uses a URL query parameter to obtain the hash key. Refer to the Envoy documentation for more information. |
Values
- Default: None
- Data type: String containing one of the following values:
LoadBalancer{}.HashPolicies[].FieldValue
Specifies the value to hash, such as a header name, cookie name, or a URL query parameter name. You cannot specify the FieldValue
parameter if SourceIP
is also configured.
Values
- Default: None
- Data type: String
LoadBalancer{}.HashPolicies[].CookieConfig
Specifies additional configuration options for the cookie
hash policy type. This field causes Envoy to generate a cookie for a client on its first request.
Values
Default: None
Data type: Map that can contain the following parameters:
Parameter Description Data type Default Session
Directs Consul to generate a session cookie with no expiration. Boolean false
TTL
Specifies the TTL for generated cookies. Cannot be specified for session cookies. String 0s
Path
Specifies the path to set for the cookie. String None
LoadBalancer{}.HashPolicies[].SourceIP
Determines if the hash type should be source IP address. You cannot specify SourceIP
if Field
or FieldValue
are configured.
Values
- Default:
false
- Data type: Boolean
LoadBalancer{}.HashPolicies[].Terminal
Determines if Consul should stop computing the hash when multiple hash policies are present. If a hash is computed when a terminal policy is evaluated, then that hash is used and subsequent hash policies are ignored.
Values
- Default:
false
- Data type: Boolean
Examples
The following examples demonstrate common service resolver configuration patterns for specific use cases.
Filter on service version
The following example creates two subsets of the web
service and assigns service instances to subsets based on each instance's version metadata. It also defines v1
as the default subset.
Kind = "service-resolver"
Name = "web"
DefaultSubset = "v1"
Subsets = {
v1 = {
Filter = "Service.Meta.version == v1"
}
v2 = {
Filter = "Service.Meta.version == v2"
}
}
Resolve virtual upstreams
The following example uses the Redirect
parameter to expose a set of services to another datacenter as a virtual service.
Kind = "service-resolver"
Name = "web-dc2"
Redirect {
Service = "web"
Datacenter = "dc2"
}
Service failover
The following example enables failover for subset v2
to dc2
. All other subsets attempt failover to dc3
, and when it is not available, to dc4
:
Kind = "service-resolver"
Name = "web"
ConnectTimeout = "15s"
Failover = {
v2 = {
Datacenters = ["dc2"]
}
"*" = {
Datacenters = ["dc3", "dc4"]
}
}
Failover targets for federation and cluster peering
The following example enables failover to target a WAN federated datacenter for all service subsets. If the connection to dc3
times out after 15 seconds, the service failover targets the peer with the establish cluster peering connection instead.
Kind = "service-resolver"
Name = "web"
ConnectTimeout = "15s"
Failover = {
"*" = {
Targets = [
{Datacenter = "dc3"},
{Peer = "peer-01"}
]
}
}
Failover for all service subsets Enterprise
The following example enables failover to the secondary
namespace in another datacenter for all service subsets.
Kind = "service-resolver"
Name = "product-api"
Namespace = "primary"
ConnectTimeout = "0s"
Failover = {
"*" = {
Datacenters = ["dc2"]
Namespace = "secondary"
}
}
Failover to a namespace Enterprise
The following example enables failover to a different namespace in the same datacenter.
Kind = "service-resolver"
Name = "product-api"
Namespace = "primary"
ConnectTimeout = "0s"
Failover = {
"*" = {
Service = "product-api-backup"
Namespace = "secondary"
}
}
Consistent load balancing
The following example applies a Maglev load balancing policy for requests with an x-user-id
header:
Kind = "service-resolver"
Name = "web"
LoadBalancer = {
Policy = "maglev"
HashPolicies = [
{
Field = "header"
FieldValue = "x-user-id"
}
]
}