# Prometheus CPU-Usage missing with Lens 2025.10.230725-latest **URL:** \ **Category:** Lens Support & Troubleshooting\ **Tags:** lens-desktop\ **Created:** [November 19, 2025, 9:25am UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631 "2025-11-19T09:25:57Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![skroehs](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/skroehs/32/1675_2.png) [@skroehs](https://forums.k8slens.dev/u/skroehs)\ **Post date:** [November 19, 2025, 9:25am UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/1 "2025-11-19T09:25:57Z")
After updating to newest Lens-version (Lens 2025.10.230725-latest) the overview of Pods no longer shows the CPU-Usage. ![lens_cpu_usage_notavailable](https://discuss-k8slens-dev.s3.dualstack.eu-west-1.amazonaws.com/original/2X/9/9b8605b530313ccfedc94bd66f9cedac6dfdbae3.png) All other metrics seem to work as before. In Metrics settings for the clusters i use Metrics Source: Prometheus Prometheus: Helm Prometheus Service Address set explicitly to “namespace/service:port” active checkbox “Filter empty containers from the Prometheus queries. Try changing this to fix missing metrics.” Prometheus version used: 3.1.0 Anyone else seeing this behaviour? ---
**Author:** ![okoshevka](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/okoshevka/32/999_2.png) [@okoshevka](https://forums.k8slens.dev/u/okoshevka)\ **Post date:** [November 19, 2025, 11:04am UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/2 "2025-11-19T11:04:15Z")
Dear @skroehs Thank you for reaching out to Lens Support! We will be sure to update our development team about your request. Warm regards, Oleksandr from Lens Support ---
**Author:** ![phorsmalahti](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/phorsmalahti/32/560_2.png) [@phorsmalahti](https://forums.k8slens.dev/u/phorsmalahti)\ **Post date:** [November 19, 2025, 11:17am UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/3 "2025-11-19T11:17:27Z")
I suspect this is related to the change “Fetch Pod CPU/Memory metrics from the selected Metrics Source.”. If you select “Kubernetes Metrics Server” as the Metrics source (temporarily for testing), does that make the CPU metric visible? Assuming Metrics Server is installed to the cluster. ---
**Author:** ![skroehs](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/skroehs/32/1675_2.png) [@skroehs](https://forums.k8slens.dev/u/skroehs)\ **Post date:** [November 19, 2025, 12:14pm UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/4 "2025-11-19T12:14:29Z")
Yes, when selecting “Kubernetes Metrics Server” as the Metrics source the CPU-Usage is visible again: ![lens_cpu_usage_metricsserver](https://discuss-k8slens-dev.s3.dualstack.eu-west-1.amazonaws.com/original/2X/b/b99d1a1df3858a6c2ac02022a61f7de1baa02850.png) ---
**Author:** ![phorsmalahti](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/phorsmalahti/32/560_2.png) [@phorsmalahti](https://forums.k8slens.dev/u/phorsmalahti)\ **Post date:** [November 19, 2025, 12:50pm UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/5 "2025-11-19T12:50:46Z")
So, looks like CPU metrics can’t be fetched from Prometheus. But I can’t tell why that would be. Maybe try using a different query type (Prometheus option), or investigate/debug the data for CPU more. ---
**Author:** ![skroehs](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/skroehs/32/1675_2.png) [@skroehs](https://forums.k8slens.dev/u/skroehs)\ **Post date:** [March 24, 2026, 3:56pm UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/6 "2026-03-24T15:56:44Z")
Problem in my case was the scraping interval configured for prometheus which was globally set to 1m. In this case this query for cpu-usage will return no results, since it does need at least 2 data points: sum(rate(container\_cpu\_usage\_seconds\_total{container!=“”, container!=“POD”, namespace=“”}[1m])) by (pod) Reducing the scrape interval for the cadvisor job (responsible for container\_cpu\_usage\_seconds\_total) solved the issue. kubernetes-nodes-cadvisor: scrape\_interval: 15s ---
**Author:** ![crothmann](https://forums.k8slens.dev/user_avatar/forums.k8slens.dev/crothmann/32/1544_2.png) [@crothmann](https://forums.k8slens.dev/u/crothmann)\ **Post date:** [April 15, 2026, 7:11am UTC](https://forums.k8slens.dev/t/prometheus-cpu-usage-missing-with-lens-2025-10-230725-latest/6631/7 "2026-04-15T07:11:56Z")
Hi @skroehs, sorry for the delayed response – and thanks a lot for taking the time to come back and share your findings. We’ve seen similar Prometheus query edge cases before, so this is really valuable for the community.