Yo Motherboard

Why is Traefik Using So Much CPU and Memory in My Homelab?

March 17, 2026 4 answers
Asked by CloudySkies99
I'm running a personal homelab cluster with RKE2 1.34, using Cilium and Ingress Nginx. Since Ingress Nginx is deprecated, I'm trying to transition to Traefik, but I'm seeing some shocking resource usage—around 90% CPU across all four cores and over 10GB of RAM. My setup isn't too large; I have about 10 ingresses and roughly 20 pods. I've removed and re-added Traefik to the cluster and even tried trial and error without success. Debugging with LLMs hasn't helped either. Here's the Traefik configuration I'm currently using. I've also attempted upgrading to RKE2 1.35, but the issue persists. Any advice on how to resolve this?
CPU Kubernetes memory resource usage System Operations Traefik troubleshooting

Answers

TechieNerd42
It sounds like that CPU and memory usage is way off for your setup! A couple of things to consider: you've likely got too many providers active—CRD, Ingress, and nginx migration could be causing the overload. Try disabling the kubernetesIngressNginx and see if it helps. Also, check those logs; it might be stuck in a resync loop, which could explain the high resource consumption.
Reply to this answer
SysAdminGuru34
From what I've seen, Traefik can be a bit aggressive with memory usage, often consuming whatever is available to keep GC loads low. Try limiting its resource allocation—you might see a significant decrease in memory usage without impacting performance.
Reply to this answer
DevOpsWiz77
If you want to dig deeper into what's causing the spike, profiling could help pinpoint the trouble spots. But how do you even go about profiling something like this? Would toggling options in the logs work, or do you have specific tools in mind?
Reply to this answer
MatrixMaster21
UPDATE: I figured out what was causing the strain! By increasing the verbosity for Traefik, I noticed a lot of 404 and 502 errors linked to my Matrix Synapse ingresses. Once I deleted those, the resource usage dropped drastically. The issue stemmed from the handling of targetPort for external name services—it needed adjustments to work properly with Traefik. After fixing that and re-applying the correct service settings, everything is back to normal (mostly, aside from a few TLS errors due to missing annotations).
Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions