feat: Show pod requests and limits on component/service metrics views

[maciaszczykm]

Test Plan

Test environment: https://console.your-env.onplural.sh/

Checklist

• I have added a meaningful title and summary to convey the impact of this PR to a user.
• If required, I have updated the Plural documentation accordingly.
• I have added tests to cover my changes.
• I have deployed the agent to a test environment and verified that it works as expected (required only when changing agent code).

Plural Flow: console

[linear]

PROD-5130

[soffi-ai]

Soffi AI Summary

This PR surfaces Kubernetes pod resource requests and limits alongside existing CPU/memory usage metrics in the component, service, and cluster metrics views. Previously, users could only see raw utilization; now they can compare actual usage against configured requests and limits directly in the UI charts.

The change is full-stack: new Prometheus queries for requests/limits are added to the Elixir metrics backend, the GraphQL schema is extended with cpuRequests, cpuLimits, memRequests, memLimits, podCpuRequests, podCpuLimits, podMemRequests, and podMemLimits fields on both ServiceComponentMetrics and KubernetesControllerMetrics, the Go client models are regenerated, and the React metric graph components are updated to fetch and render these new series. A Kubernetes spec fallback is also implemented so that if Prometheus data is unavailable, the UI can still display the limits/requests values derived from the pod spec.

Changes

Pod resource requests and limits in metrics views

• Adds Prometheus queries for CPU/memory requests and limits (both controller-level and pod-level) to the Elixir observability metrics module, and exposes the new fields in the GraphQL schema for ServiceComponentMetrics and KubernetesControllerMetrics. (22fc4de)
• Wires the new requests/limits metric fields through the GraphQL fragments and React components (ServiceMetrics, ComponentMetrics, KubernetesMetrics, ResourceMetricsGraphs), rendering them as additional series on the existing CPU/memory charts with a fallback to Kubernetes pod spec values when Prometheus data is absent. Regenerates the Go client models and TypeScript GraphQL types. (7ac6bd5)
• Merge commit integrating upstream master changes into the feature branch. (354783a)

---

Updated: 2026-09-16 14:36 UTC

Deploy in Soffi

[maciaszczykm]

@greptileai

[greptile-apps]

Confidence Score: 5/5

The PR appears safe to merge, with no outstanding correctness or repository-rule issues identified.

Summary

The PR extends service, component, and Kubernetes metrics views to display CPU and memory requests and limits alongside usage.

• Adds Prometheus queries and GraphQL fields for aggregate and per-pod reservations.
• Keeps backend schema, frontend operations, generated client types, and persisted queries synchronized.
• Reuses shared resource graph rendering across metrics views.
• Adds distinguishable backend assertions for all new metric mappings.
• Updates empty-state detection so reservation-only data remains visible.

_{Reviews (3) · Last reviewed commit: "yarn fix"}

[maciaszczykm]

@greptileai

[michaeljguarino]

can i have a screenshot of the change here? This was sort of being done already

[maciaszczykm]

@michaeljguarino Before, requests and limits were flat lines from the current pod spec. This PR gets that data from Prometheus (with the spec as fallback on component and Kubernetes graphs), so when possible it shows real values over time. It also adds these lines to service graphs and to the overall CPU/memory charts.