Prefix Utilization

[klinkeye]

Is it expected that the parent prefix with child prefixes shows 0% utilization when the child prefixes have various values of utilization above 0%?

I would have expected the parent prefix would reflect the level of utilization of the total parent prefix with respect to the amount of child utilization.

I am a network engineer, but I might be coming at this from a different direction than others using Netbox.

[jeremystretch]

Parent prefix utilization is calculated according to the total amount of their own space consumed by allocated children. You can see examples of this on the demo instance. It's impossible to say whether what you're seeing on your local instance is correct or not without more detail.

[klinkeye]

Thanks Jeremy, mine is definitely not appearing like the demo. I have a parent prefix, with child prefixes. If I mark some of those child prefixes as 100% utilized, the parent stays at 0% utilized.

Im on 3.1.8

[candlerb]

Are they in different VRFs? If so they're considered non-overlapping. Otherwise, can you give the exact set of steps to create the prefixes you're talking about?

[klinkeye]

Are they in different VRFs? If so they're considered non-overlapping. Otherwise, can you give the exact set of steps to create the prefixes you're talking about?

All in the global VRF, see screenshot

[klinkeye]

Ok so doing some trial and error. If you populate the IP addresses manually and individually within a prefix it shows the correct parent utilization, however if you select the "Mark Utilized" checkbox (treat as 100% utilized) in a child prefix's properties the parent prefix ignores that.

[candlerb]

Try changing the prefixes 10.200.0.0.16/, 10.200.200.0/24 and 10.200.255.0/24 from "Active" to "Container". See #1689 and #8091.

As I understand it:

• Prefixes marked as "Active" have their utilization counted based on the number of IP addresses contained, ignoring any child prefixes
• Prefixes marked as "Container" have their utilization counted based on the amount of address space carved into child prefixes - regardless of the utilization of those child prefixes.

So if you have say:

• Container Prefix 192.168.0.0/24
  • Active Prefix 192.168.0.0/26
    • Address 192.168.0.1/26
    • Address 192.168.0.2/26

then the utilization of the container is 25%, since 1/4 of its address space has been taken up by the child prefix. Whether the actual utilization of 192.168.0.0/26 is zero or 100% or anywhere between makes no difference; the container only has 75% remaining space for allocation of further child prefixes.

The utilization of 192.168.0.0/26 will be 2/64 (or 2/62). If you mark it as 100% utilized this shows in its own utilization, but it doesn't affect the calculation of outer prefixes. Why?

• If 192.168.0.0/24 is "Container" then its utilization is already marked at 25%, treating the whole child prefix as allocated
• If 192.168.0.0/24 is "Active" then its utilization would be 2/256 (or 2/254), based on child IP addresses only. Since it's only looking at IP addresses, it makes no difference whether the 192.168.0.0/26 prefix is marked "100% utilized" or not.

Note: for this to work properly, a "Container" prefix should only contain prefixes, and an "Active" prefix should only contain IP addresses.

You can get into problems if a "Container" prefix contains IP addresses directly - e.g. for allocating loopbacks. Those problems can be fixed by using "Reserved" instead of "Container": see #5925. (Anything which is not a "Container" is treated the same for the purposes of calculating utilization)

[vladsol]

It would be great if the parent prefix reflects the usage of the child prefix if it is marked as 100% utilized.

[mrkeathley]

Was there an issue generated for this? I was working on a fix internally but didn't want to replicate another solution.