Get desktop application:
View/edit binary Protocol Buffers messages
Lookup returns a target for a single key.
Target type allows the client to specify what kind of target format it would like from RLS to allow it to find the regional server, e.g. "grpc".
Reason for making this request.
For REASON_STALE, the header_data from the stale response, if any.
Map of key values extracted via key builders for the gRPC or HTTP request.
Application-specific optional extensions.
Prioritized list (best one first) of addressable entities to use for routing, using syntax requested by the request target_type. The targets will be tried in order until a healthy one is found.
Optional header value to pass along to AFE in the X-Google-RLS-Data header. Cached with "target" and sent with all requests that match the request key. Allows the RLS to pass its work product to the eventual target.
Application-specific optional extensions.
A GrpcKeyBuilder applies to a given gRPC service, name, and headers.
Used in:
Extract keys from all listed headers. For gRPC, it is an error to specify "required_match" on the NameMatcher protos.
You can optionally set one or more specific key/value pairs to be added to the key_map. This can be useful to identify which builder built the key, for example if you are suppressing the actual method, but need to separately cache and request all the matched methods.
If you wish to include the host, service, or method names as keys in the generated RouteLookupRequest, specify key names to use in the extra_keys submessage. If a key name is empty, no key will be set for that value. If this submessage is specified, the normal host/path fields will be left unset in the RouteLookupRequest. We are deprecating host/path in the RouteLookupRequest, so services should migrate to the ExtraKeys approach.
Used in:
To match, one of the given Name fields must match; the service and method fields are specified as fixed strings. The service name is required and includes the proto package name. The method name may be omitted, in which case any method on the given service is matched.
Used in:
An HttpKeyBuilder applies to a given HTTP URL and headers. Path and host patterns use the matching syntax from gRPC transcoding to extract named key/value pairs from the path and host components of the URL: https://github.com/googleapis/googleapis/blob/master/google/api/http.proto It is invalid to specify the same key name in multiple places in a pattern. For a service where the project id can be expressed either as a subdomain or in the path, separate HttpKeyBuilders must be used: host_pattern: 'example.com' path_pattern: '/{id}/{object}/**' host_pattern: '{id}.example.com' path_pattern: '/{object}/**' If the host is exactly 'example.com', the first path segment will be used as the id and the second segment as the object. If the host has a subdomain, the subdomain will be used as the id and the first segment as the object. If neither pattern matches, no keys will be extracted.
Used in:
host_pattern is an ordered list of host template patterns for the desired value. If any host_pattern values are specified, then at least one must match, and the last one wins and sets any specified variables. A host consists of labels separated by dots. Each label is matched against the label in the pattern as follows: - "*": Matches any single label. - "**": Matches zero or more labels (first or last part of host only). - "{<name>=...}": One or more label capture, where "..." can be any template that does not include a capture. - "{<name>}": A single label capture. Identical to {<name>=*}. Examples: - "example.com": Only applies to the exact host example.com. - "*.example.com": Matches subdomains of example.com. - "**.example.com": matches example.com, and all levels of subdomains. - "{project}.example.com": Extracts the third level subdomain. - "{project=**}.example.com": Extracts the third level+ subdomains. - "{project=**}": Extracts the entire host.
path_pattern is an ordered list of path template patterns for the desired value. If any path_pattern values are specified, then at least one must match, and the last one wins and sets any specified variables. A path consists of segments separated by slashes. Each segment is matched against the segment in the pattern as follows: - "*": Matches any single segment. - "**": Matches zero or more segments (first or last part of path only). - "{<name>=...}": One or more segment capture, where "..." can be any template that does not include a capture. - "{<name>}": A single segment capture. Identical to {<name>=*}. A custom method may also be specified by appending ":" and the custom method name or "*" to indicate any custom method (including no custom method). For example, "/*/projects/{project_id}/**:*" extracts `{project_id}` for any version, resource and custom method that includes it. By default, any custom method will be matched. Examples: - "/v1/{name=messages/*}": extracts a name like "messages/12345". - "/v1/messages/{message_id}": extracts a message_id like "12345". - "/v1/users/{user_id}/messages/{message_id}": extracts two key values.
List of query parameter names to try to match. For example: ["parent", "name", "resource.name"] We extract all the specified query_parameters (case-sensitively). If any are marked as "required_match" and are not present, this keybuilder fails to match. If a given parameter appears multiple times (?foo=a&foo=b) we will report it as a comma-separated string (foo=a,b).
List of headers to try to match. We extract all the specified header values (case-insensitively). If any are marked as "required_match" and are not present, this keybuilder fails to match. If a given header appears multiple times in the request we will report it as a comma-separated string, in standard HTTP fashion.
You can optionally set one or more specific key/value pairs to be added to the key_map. This can be useful to identify which builder built the key, for example if you are suppressing a lot of information from the URL, but need to separately cache and request URLs with that content.
If specified, the HTTP method/verb will be extracted under this key name.
Extract a key based on a given name (e.g. header name or query parameter name). The name must match one of the names listed in the "name" field. If the "required_match" field is true, one of the specified names must be present for the keybuilder to match.
Used in: ,
The name that will be used in the RLS key_map to refer to this value. If required_match is true, you may omit this field or set it to an empty string, in which case the matcher will require a match, but won't update the key_map.
Ordered list of names (headers or query parameter names) that can supply this value; the first one with a non-empty value is used.
If true, make this extraction required; the key builder will not match if no value is found.
RouteLookupClusterSpecifier is used in xDS to represent a cluster specifier plugin for RLS.
The RLS config for this cluster specifier plugin instance.
Used in: ,
Ordered specifications for constructing keys for HTTP requests. Last match wins. If no HttpKeyBuilder matches, an empty key_map will be sent to the lookup service; it should likely reply with a global default route and raise an alert.
Unordered specifications for constructing keys for gRPC requests. All GrpcKeyBuilders on this list must have unique "name" fields so that the client is free to prebuild a hash map keyed by name. If no GrpcKeyBuilder matches, an empty key_map will be sent to the lookup service; it should likely reply with a global default route and raise an alert.
The name of the lookup service as a gRPC URI. Typically, this will be a subdomain of the target, such as "lookup.datastore.googleapis.com".
Configure a timeout value for lookup service requests. Defaults to 10 seconds if not specified.
How long are responses valid for (like HTTP Cache-Control). If omitted or zero, the longest valid cache time is used. This value is clamped to 5 minutes to avoid unflushable bad responses, unless stale_age is specified.
After a response has been in the client cache for this amount of time and is re-requested, start an asynchronous RPC to re-validate it. This value should be less than max_age by at least the length of a typical RTT to the Route Lookup Service to fully mask the RTT latency. If omitted, keys are only re-requested after they have expired. This value is clamped to 5 minutes.
Rough indicator of amount of memory to use for the client cache. Some of the data structure overhead is not accounted for, so actual memory consumed will be somewhat greater than this value. If this field is omitted or set to zero, a client default will be used. The value may be capped to a lower amount based on client configuration.
This is a list of all the possible targets that can be returned by the lookup service. If a target not on this list is returned, it will be treated the same as an unhealthy target.
This value provides a default target to use if needed. If set, it will be used if RLS returns an error, times out, or returns an invalid response. Note that requests can be routed only to a subdomain of the original target, e.g. "us_east_1.cloudbigtable.googleapis.com".
Possible reasons for making a request.
Used in:
Unused
No data available in local cache
Data in local cache is stale