How should print tablesize be interpreted for dynamic BGP neighbors using neighbor-range / neighbor-group?

Hello everyone,

I would like to clarify how FortiGate handles dynamic BGP neighbors in relation to the print tablesize output.

In our environment, we use a hub-and-spoke topology. On the hub, BGP peers are not configured statically one by one under config neighbor. Instead, we use:

• neighbor-group
• neighbor-range

This allows spoke peers to establish BGP sessions dynamically.

When running print tablesize, we see:

router.bgp:neighbor: 0 0 1000

My doubt is about the meaning of this value in a dynamic BGP scenario.

I understand that max-neighbor-num under neighbor-range is a per-range configurable control, not necessarily the total platform limit. For example, the CLI allows:

set max-neighbor-num <1-1000> or 0 and documentation mentions that 0 is the default special value.

Because of that, I would like to understand the following:

1. Does router.bgp:neighbor in print tablesize apply only to statically configured neighbors, or does it also include dynamic neighbors created through neighbor-range?
2. If max-neighbor-num is set to 0, should this be interpreted as:
   • no explicit per-range limit, and
   • only the platform/runtime capacity still applies?
3. Is there any official Fortinet reference that defines the maximum supported number of dynamic BGP peers/sessions for a given model?
4. In general, for this case, which value should be used for capacity planning:
   • print tablesize
   • datasheet/platform scale
   • or another BGP-specific limit?

I am asking because in other features such as IPsec, print tablesize seems to describe configuration object limits, while actual feature scale may depend on runtime/platform behavior.

Any clarification, official reference, or field experience would be appreciated.

Thanks.