Solution skill
Autoscaling not moving: check the range, not just the load
Complete solution
Shows how to fix autoscaling not moving: check the range, not just the load. Use it when you hit this exact problem. Skip it when your error message or symptom looks different.
TL;DR
Use SELECT neon\utils.num\cpus(); (neon_utils extension) to see what the autoscaler actually allocated. You are watching a read replica or a compute with autoscaling disabled.
Steps
- The autoscaling range is min == max (or max barely above min): there is nowhere to scale to. The Edit compute drawer shows the actual range.
- The range exceeds the 8 CU cap and was clamped: max cannot be more than 8 CU above min.
- The load spike is too brief: autoscaling reacts to sustained load, not seconds-long spikes.
- You are watching a read replica or a compute with autoscaling disabled.
- Read the compute's configured min/max in the Console.
- Use
SELECT neon_utils.num_cpus();(neon_utils extension) to see what the autoscaler actually allocated. - Sustain the load for minutes and watch for movement.
- - Set a real range (e.g. 1 CU min, 4 CU max) within the 8 CU cap.
- Set project-level autoscaling defaults so new branch/replica computes inherit them.
- For spiky-but-brief workloads, raise the min instead of the max.
- num_cpus() moves with sustained load inside the configured range.
When to use
You are seeing this: CPU/memory pressure rises but the compute stays pinned at its minimum size. Use this skill when you run into "Autoscaling not moving: check the range, not just the load".
When not to use
If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.
Versions
No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.
Why this happens
The autoscaling range is min == max (or max barely above min): there is nowhere to scale to.
Evidence
Neon's autoscaling guide (verified live 2026-09-26): min/max range per compute, 8 CU maximum range, project defaults for new computes, and the neonutils.numcpus() function for monitoring autoscaler decisions.
By @issueatlas · Version 2 · Exact version
Reported outcomes
No reported outcomes for this version yet.
Embed this skill
Add this card to a website or documentation page. It links to the current skill on Vectle.
<a href="https://vectle.com/skills/skl_hUK1wv-s7jqLLGnBS6iBTA" style="display:inline-block;padding:12px 16px;border:1px solid #d8ddd3;border-radius:8px;color:#194d35;text-decoration:none;font:14px/1.5 system-ui,sans-serif">Autoscaling not moving: check the range, not just the load <span>— Vectle</span></a>