Still an issue in 42.20, Basically blank project, over 7.0 and Red.
1 Like
With the 42.20 update my memory usage reading have gone up. I was getting numbers at 5.4 to 5.6 on average. Now I’m consistently getting numbers up around 5.9 on average on a blank map.
1 Like
Thank you for all the engagement with this issue. On the frame times, we have improvements scheduled for this year to better communicate that it’s Fortnite energy saving kicking in when the game client is out of focus.
For the high memory consumption on an empty map exceeding the threshold, we’re making and testing improvements. The files you’ve provided us have been a great help, and we’ve found an internal repro. We’re very focused on resolving this and will keep you updated on the improvements we make 
We’ll keep active in this post
4 Likes
@Aborup That is encouraging. Thank you for the update. I see this feature as being vital/extremely useful for me and my game, and I haven’t been able to rely on it at all yet.
Hi everyone,
Getting the same issue here on UEFN v41.20 (latest version).
Spatial Profiler is reporting Memory usage around 6.7 GiB out of the 7.2 GiB critical threshold, even on a map that isn’t heavy content-wise. Screenshot attached.
I’ve also noticed the number is very inconsistent across sessions. I checked Windows Task Manager during live editing and sometimes the two Fortnite processes combined only add up to about 4.6 GB total RAM, well under the threshold. But on other days, with no changes made to the map at all, the Spatial Profiler reading climbs up to around 7.3 GB and goes over the critical limit. It fluctuates from one day to another without any modification on my end, so I can’t tell what’s actually driving it.
Does anyone know what could be causing this discrepancy, or what might explain the fluctuation? Is this a known issue with how the Spatial Profiler calculates memory, and is there a recommended way to get a more reliable reading?
Thanks in advance!
1 Like
Hi everyone,
I wanted to share some additional data on the memory reporting issue discussed in this thread, since I’m seeing significant inconsistency across sessions with no changes made to my project in between.
My island is currently at roughly 2% of the total map size/complexity budget, so there is no realistic explanation for memory usage anywhere near the critical threshold, let alone several times over it.
Logged across different sessions, same project, zero edits made between them:
- Session 1: ~4.6 GB (cross-checked with Task Manager, which roughly matched the Spatial Profiler reading that time)
- Session 2: ~7.1-7.3 GiB on the Spatial Profiler, again roughly matching real system RAM usage
- Session 3 (most recent): 17.1 GiB out of the 7.2 GiB budget, with frame time also spiking to 36.6 ms / 16.7 ms — no changes made to the project since the previous session
A Memory Snapshot of my actual content only accounts for a few hundred MiB at most (textures, static meshes, and materials combined), so whatever is driving these swings clearly isn’t coming from my content.
For reference, my system specs are:
- CPU: AMD Ryzen 7 5800X (8-core / 16 threads, ~3.8GHz)
- RAM: 32 GB
- GPU: NVIDIA GeForce RTX 5060, 8 GB VRAM
- Motherboard: MSI MS-7C95
- OS: Windows 11 Home 64-bit (build 26200)
Given that my total system RAM is 32 GB, a reported 17.1 GiB for this process alone doesn’t align with actual system behavior during that session.
This level of variance (4.6 GB to 17.1 GB on the same near-empty project) makes the memory metric unreliable for judging whether an island is within budget, and I’m unable to tell whether this is a leak that accumulates over a session, a caching issue, or something else.
Would it be possible for someone on the team to take a look at this? I’m happy to provide my Memory Snapshot file and logs (UEFN and Fortnite) if that would help investigate further.
Thank you.
1 Like
Seeing same numbers, similar spec
FORT-1129055’s status has changed to ‘Ready for QA’. A member of the QA department is investigating the issue.
Thanks for the additional evidence. We have found that the reported memory usage is not always accurate with the actual consumed memory. We’re testing on our internal workstation that reproduces inflated memory numbers to ensure the UI is reporting the correct memory usage. We’re QA’ing that now to measure if the improvement is satisfactory.
I think some of the confusion is were thinking this is the memory of the island but it seems to be the memory of the default graphics card ? this is always max value/value - Sidenote it doesn’t reflect your actual graphics card gpu memory if this is how I’m understanding it to be as i have 16gb video memory yet always shows about 8gb.
Hope this gets fixed though
It’s showing the memory usage of the Fortnite client running your island. It’s not exclusively memory used by your island - it’s the whole game while running your content. The Memory Snapshot tool can show you how your island’s assets, and the assets it references, contribute to that memory use.
@R_LIX2 For that session 3 number, did you confirm what Task Manager was showing? I think the committed memory that was being reported in your session 1 and 2 may be explainable variance (driver and GPU upload memory and friends). Going from 7 GiB to 17 GiB between runs with no change is suspect.
To compare the memory figure with task manager:
- Open Task Manager and go to the Details tab.
- Right-click any column header, choose Select columns, and turn on Commit size, Memory (private working set) and Working set (memory).
- How do those numbers compare to what UEFN is showing you?