Introduce non-AFI-SAFI aggregate prefix limit at global, neighbor, and peer-group levels

[mukherso]

Files changed:

• (M) release/models/bgp/openconfig-bgp-common-multiprotocol.yang
• (M) release/models/bgp/openconfig-bgp-common-structure.yang
• (M) release/models/bgp/openconfig-bgp-global.yang
• (M) release/models/bgp/openconfig-bgp-neighbor.yang
• (M) release/models/bgp/openconfig-bgp-peer-group.yang

Change Scope:

This change introduces a non-AFI-SAFI aggregate prefix-limit for accepted routes and prefix-limit-received for received routes at the BGP global, neighbor, and peer-group levels, and renames one internal grouping to reflect its broadened scope.

1. Grouping rename (openconfig-bgp-common-multiprotocol.yang):

• bgp-common-mp-all-afi-safi-common-prefix-limit-config → bgp-common-prefix-limit-config

The original name implied this grouping was scoped exclusively to AFI-SAFI contexts. Since it is now reused outside that context, the name has been shortened accordingly. bgp-common-mp-all-afi-safi-common-prefix-limit-state is unchanged as it remains AFI-SAFI specific. All existing uses statements within the AFI-SAFI prefix-limit containers are updated accordingly. There is no functional change to the existing per-AFI-SAFI prefix-limit behaviour.

2. New structural grouping (openconfig-bgp-common-structure.yang):

A new grouping bgp-common-structure-prefix-limit is added, following the existing OpenConfig config/state container pattern. It is applied at the global, neighbor, and peer-group levels via openconfig-bgp-global.yang, openconfig-bgp-neighbor.yang, and openconfig-bgp-peer-group.yang respectively.

This allows operators to set a single aggregate prefix-limit across all address families for a session, complementing the existing per-AFI-SAFI prefix-limit. Note that prefix-limit-exceeded is intentionally not included in the state container at these levels, as exceeded-state tracking is only modelled at the AFI-SAFI granularity.

Relevant pyang tree output:

module: openconfig-bgp
    +--rw bgp
       +--rw global
       |  +--rw prefix-limit-received
       |     +--rw config
       |     |  +--rw max-prefixes?            uint32
       |     |  +--rw prevent-teardown?        boolean
       |     |  +--rw warning-threshold-pct?   oc-types:percentage
       |     +--ro state
       |        +--ro max-prefixes?            uint32
       |        +--ro prevent-teardown?        boolean
       |        +--ro warning-threshold-pct?   oc-types:percentage
       +--rw neighbors
       |  +--rw neighbor* [neighbor-address]
       |     +--rw prefix-limit
       |        +--rw config
       |        |  +--rw max-prefixes?            uint32
       |        |  +--rw prevent-teardown?        boolean
       |        |  +--rw warning-threshold-pct?   oc-types:percentage
       |        +--ro state
       |           +--ro max-prefixes?            uint32
       |           +--ro prevent-teardown?        boolean
       |           +--ro warning-threshold-pct?   oc-types:percentage
       +--rw peer-groups
          +--rw peer-group* [peer-group-name]
             +--rw prefix-limit
                +--rw config
                |  +--rw max-prefixes?            uint32
                |  +--rw prevent-teardown?        boolean
                |  +--rw warning-threshold-pct?   oc-types:percentage
                +--ro state
                   +--ro max-prefixes?            uint32
                   +--ro prevent-teardown?        boolean
                   +--ro warning-threshold-pct?   oc-types:percentage

Platform Implementations

Arista EOS User Guide implements BGP prefix-limit-received using the neighbor maximum-routes command, which enforces a per-session prefix threshold and may terminate the BGP session upon exceedance.

[gemini-code-assist]

Code Review

This pull request implements an aggregate prefix limit across BGP global, neighbor, and peer-group contexts by generalizing existing groupings and introducing a new structural grouping. The review feedback suggests correcting the container description to use "accepted" instead of "received" to align with OpenConfig conventions and recommends adding a state leaf to track when the aggregate limit is exceeded for better operational visibility.