xDS REST and gRPC protocol

    The simplest approach to delivering dynamic configuration is to place it at a well known path specified in the ConfigSource. Envoy will use inotify (kqueue on macOS) to monitor the file for changes and parse the proto in the file on update. Binary protobufs, JSON, YAML and proto text are supported formats for the DiscoveryResponse.

    There is no mechanism available for filesystem subscriptions to ACK/NACK updates beyond stats counters and logs. The last valid configuration for an xDS API will continue to apply if an configuration update rejection occurs.

    A gRPC can be specified independently for each xDS API, pointing at an upstream cluster corresponding to a management server. This will initiate an independent bidirectional gRPC stream for each xDS resource type, potentially to distinct management servers. API delivery is eventually consistent. See ` below for situations in which explicit control of sequencing is required.

    Type URLs

    Each xDS API is concerned with resources of a given type. There is a 1:1 correspondence between an xDS API and a resource type. That is:

    The concept of appears below, and takes the form type.googleapis.com/, e.g. type.googleapis.com/envoy.api.v2.Cluster for CDS. In various requests from Envoy and responses by the management server, the resource type URL is stated.

    ACK/NACK and versioning

    Each stream begins with a from Envoy, specifying the list of resources to subscribe to, the type URL corresponding to the subscribed resources, the node identifier and an empty version_info. An example EDS request might be:

    The management server may reply either immediately or when the requested resources are available with a , e.g.:

    After processing the DiscoveryResponse, Envoy will send a new request on the stream, specifying the last version successfully applied and the nonce provided by the management server. If the update was successfully applied, the will be X, as indicated in the sequence diagram:

    In this sequence diagram, and below, the following format is used to abbreviate messages:

    • DiscoveryRequest: (V=version_info,R=resource_names,N=response_nonce,T=type_url)
    • DiscoveryResponse: (V=version_info,R=resources,N=nonce,T=type_url)

    The version provides Envoy and the management server a shared notion of the currently applied configuration, as well as a mechanism to ACK/NACK configuration updates. If Envoy had instead rejected configuration update X, it would reply with error_detail populated and its previous version, which in this case was the empty initial version. The has more details around the exact error message populated in the message field:

    No version update after NACK

    Later, an API update may succeed at a new version Y:

    Each stream has its own notion of versioning, there is no shared versioning across resource types. When ADS is not used, even each resource of a given resource type may have a distinct version, since the Envoy API allows distinct EDS/RDS resources to point at different ConfigSources.

    When to send an update

    The management server should only send updates to the Envoy client when the resources in the DiscoveryResponse have changed. Envoy replies to any with a DiscoveryRequest containing the ACK/NACK immediately after it has been either accepted or rejected. If the management server provides the same set of resources rather than waiting for a change to occur, it will cause Envoy and the management server to spin and have a severe performance impact.

    Within a stream, new supersede any prior DiscoveryRequests having the same resource type. This means that the management server only needs to respond to the latest on each stream for any given resource type.

    Resource hints

    LDS/CDS resource hints will always be empty and it is expected that the management server will provide the complete state of the LDS/CDS resources in each response. An absent Listener or Cluster will be deleted.

    For EDS/RDS, the management server does not need to supply every requested resource and may also supply additional, unrequested resources. is only a hint. Envoy will silently ignore any superfluous resources. When a requested resource is missing in a RDS or EDS update, Envoy will retain the last known value for this resource except in the case where the Cluster or Listener is being warmed. See section below on the expectations during warming. The management server may be able to infer all the required EDS/RDS resources from the node identification in the , in which case this hint may be discarded. An empty EDS/RDS DiscoveryResponse is effectively a nop from the perspective of the respective resources in the Envoy.

    When a Listener or Cluster is deleted, its corresponding EDS and RDS resources are also deleted inside the Envoy instance. In order for EDS resources to be known or tracked by Envoy, there must exist an applied Cluster definition (e.g. sourced via CDS). A similar relationship exists between RDS and Listeners (e.g. sourced via LDS).

    For EDS/RDS, Envoy may either generate a distinct stream for each resource of a given type (e.g. if each has its own distinct upstream cluster for a management server), or may combine together multiple resource requests for a given resource type when they are destined for the same management server. While this is left to implementation specifics, management servers should be capable of handling one or more resource_names for a given resource type in each request. Both sequence diagrams below are valid for fetching two EDS resources {foo, bar}:

    Multiple EDS requests on the same stream

    Resource updates

    As discussed above, Envoy may update the list of resource_names it presents to the management server in each that ACK/NACKs a specific DiscoveryResponse. In addition, Envoy may later issue additional at a given version_info to update the management server with new resource hints. For example, if Envoy is at EDS version X and knows only about cluster , but then receives a CDS update and learns about in addition, it may issue an additional for X with {foo,bar} as resource_names.

    CDS response leads to EDS resource hint update

    There is a race condition that may arise here; if after a resource hint update is issued by Envoy at X, but before the management server processes the update it replies with a new version Y, the resource hint update may be interpreted as a rejection of Y by presenting an X version_info. To avoid this, the management server provides a that Envoy uses to indicate the specific each DiscoveryRequest corresponds to:

    The management server should not send a for any DiscoveryRequest that has a stale nonce. A nonce becomes stale following a newer nonce being presented to Envoy in a . A management server does not need to send an update until it determines a new version is available. Earlier requests at a version then also become stale. It may process multiple DiscoveryRequests at a version until a new version is ready.

    Requests become stale

    An implication of the above resource update sequencing is that Envoy does not expect a for every DiscoveryRequests it issues.

    and Listeners go through warming before they can serve requests. This process happens both during and when the Cluster or Listener is updated. Warming of Cluster is completed only when a ClusterLoadAssignment response is supplied by management server. Similarly, warming of Listener is completed only when a RouteConfiguration is supplied by management server if the listener refers to an RDS configuration. Management server is expected to provide the EDS/RDS updates during warming. If management server does not provide EDS/RDS responses, Envoy will not initialize itself during the initialization phase and the updates sent via CDS/LDS will not take effect until EDS/RDS responses are supplied.

    Eventual consistency considerations

    Since Envoy’s xDS APIs are eventually consistent, traffic may drop briefly during updates. For example, if only cluster X is known via CDS/EDS, a RouteConfiguration references cluster X and is then adjusted to cluster Y just before the CDS/EDS update providing Y, traffic will be blackholed until Y is known about by the Envoy instance.

    For some applications, a temporary drop of traffic is acceptable, retries at the client or by other Envoy sidecars will hide this drop. For other scenarios where drop can’t be tolerated, traffic drop could have been avoided by providing a CDS/EDS update with both X and Y, then the RDS update repointing from X to Y and then a CDS/EDS update dropping X.

    In general, to avoid traffic drop, sequencing of updates should follow a make before break model, wherein:

    • CDS updates (if any) must always be pushed first.
    • EDS updates (if any) must arrive after CDS updates for the respective clusters.
    • LDS updates must arrive after corresponding CDS/EDS updates.
    • RDS updates related to the newly added listeners must arrive after CDS/EDS/LDS updates.
    • VHDS updates (if any) related to the newly added RouteConfigurations must arrive after RDS updates.
    • Stale CDS clusters and related EDS endpoints (ones no longer being referenced) can then be removed.

    xDS updates can be pushed independently if no new clusters/routes/listeners are added or if it’s acceptable to temporarily drop traffic during updates. Note that in case of LDS updates, the listeners will be warmed before they receive traffic, i.e. the dependent routes are fetched through RDS if configured. Clusters are warmed when adding/removing/updating clusters. On the other hand, routes are not warmed, i.e., the management plane must ensure that clusters referenced by a route are in place, before pushing the updates for a route.

    A single ADS stream is available per Envoy instance.

    An example minimal fragment for ADS configuration is:

    Incremental xDS is a separate xDS endpoint that:

    • Allows the protocol to communicate on the wire in terms of resource/resource name deltas (“Delta xDS”). This supports the goal of scalability of xDS resources. Rather than deliver all 100k clusters when a single cluster is modified, the management server only needs to deliver the single cluster that changed.
    • Allows the Envoy to on-demand / lazily request additional resources. For example, requesting a cluster only when a request for that cluster arrives.

    An Incremental xDS session is always in the context of a gRPC bidirectional stream. This allows the xDS server to keep track of the state of xDS clients connected to it. There is no REST version of Incremental xDS yet.

    In the delta xDS wire protocol, the nonce field is required and used to pair a to a DeltaDiscoveryRequest ACK or NACK. Optionally, a response message level is present for debugging purposes only.

    DeltaDiscoveryRequest can be sent in the following situations:

    • Initial message in a xDS bidirectional gRPC stream.
    • Spontaneous from the client. This can be done to dynamically add or remove elements from the tracked resource_names set. In this case must be omitted.

    In this first example the client connects and receives a first update that it ACKs. The second update fails and the client NACKs the update. Later the xDS client spontaneously requests the “wc” resource.

    Incremental session example

    On reconnect the Incremental xDS client may tell the server of its known resources to avoid resending them over the network. Because no state is assumed to be preserved from the previous stream, the reconnecting client must provide the server with all resource names it is interested in.

    Resource names

    Resources are identified by a resource name or an alias. Aliases of a resource, if present, can be identified by the alias field in the resource of a . The resource name will be returned in the name field in the resource of a DeltaDiscoveryResponse.

    Subscribing to Resources

    The client can send either an alias or the name of a resource in the resource_names_subscribe field of a in order to subscribe to a resource. Both the names and aliases of resources should be checked in order to determine whether the entity in question has been subscribed to.

    A resource_names_subscribe field may contain resource names that the server believes the client is already subscribed to, and furthermore has the most recent versions of. However, the server must still provide those resources in the response; due to implementation details hidden from the server, the client may have “forgotten” those resources despite apparently remaining subscribed.

    Unsubscribing from Resources

    When a client loses interest in some resources, it will indicate that with the resource_names_unsubscribe field of a . As with resource_names_subscribe, these may be resource names or aliases.

    A field may contain superfluous resource names, which the server thought the client was already not subscribed to. The server must cleanly process such a request; it can simply ignore these phantom unsubscriptions.

    Synchronous (long) polling via REST endpoints is also available for the xDS singleton APIs. The above sequencing of messages is similar, except no persistent stream is maintained to the management server. It is expected that there is only a single outstanding request at any point in time, and as a result the response nonce is optional in REST-JSON. The JSON canonical transform of proto3 is used to encode and DiscoveryResponse messages. ADS is not available for REST-JSON polling.