Open
Conversation
SoWieMarkus
requested review from
PhilippMatthes,
auhlig,
juliusclausnitzer,
mblos and
umswmayj
as code owners
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
✅ Files skipped from review due to trivial changes (1)
📝 WalkthroughWalkthroughAdds Perses deployment and dashboard provisioning to Tilt: adds the Perses Helm repo and a Changes
Sequence Diagram(s)sequenceDiagram
participant Dev as Developer (Tilt)
participant Tilt as Tilt
participant Helm as Helm Chart Repo
participant K8s as Kubernetes
participant Perses as Perses Pod
participant Prom as Prometheus
Dev->>Tilt: start Tilt
Tilt->>Helm: fetch `perses/perses` chart
Helm->>K8s: create resources (Deployment, Service, ConfigMap from dashboards)
K8s->>Perses: deploy pod(s) (sidecar enabled)
Tilt->>K8s: port-forward 5080 -> 8080
Perses->>Prom: query datasource at http://prometheus-operated:9090
Dev->>Tilt: open dashboard link (localhost:5080)
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
🧹 Nitpick comments (2)
Tiltfile (2)
291-299: Consider addingcortex-prometheustoresource_deps.The Perses datasource is configured to connect to
prometheus-operated:9090. While the deployment will eventually work (Perses will retry), adding an explicit dependency would ensure proper startup ordering in the dev environment.Suggested change
helm_resource( 'cortex-perses', 'perses/perses', flags=['--values=./tools/perses/values.yaml'], port_forwards=[port_forward(5080, 8080, name='perses')], links=[link('http://localhost:5080', 'perses dashboard')], labels=['Monitoring'], - resource_deps=['perses'], + resource_deps=['perses', 'cortex-prometheus'], )🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Tiltfile` around lines 291 - 299, The Perses helm_resource ('cortex-perses') may start before Prometheus and its Service (prometheus-operated:9090) is available; update the helm_resource call for 'cortex-perses' to include 'cortex-prometheus' in the resource_deps list so Tilt waits for the Prometheus resource before creating Perses, i.e., add 'cortex-prometheus' alongside the existing 'perses' entry in resource_deps of the helm_resource invocation.
300-306: ConfigMap creation approach works but is unconventional.The piped
kubectlcommand correctly creates a ConfigMap with theperses.dev/resource=truelabel for sidecar discovery. Thewatch_fileensures Tilt rebuilds when dashboards change.One minor note: this ConfigMap won't appear as a labeled resource in Tilt's UI since there's no
k8s_resource()call for it. If you want it visible under "Monitoring", you could add:k8s_resource( new_name='cortex-perses-dashboards', objects=['cortex-perses-dashboards:ConfigMap:default'], labels=['Monitoring'], )This is optional and depends on whether you want the ConfigMap visible in Tilt's resource list.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Tiltfile` around lines 300 - 306, The generated ConfigMap created via watch_file('./tools/perses/dashboards') + k8s_yaml(...) is correct but won’t show up in Tilt’s UI; add a k8s_resource declaration to surface it in the Tilt resource list. Add a k8s_resource call that references the ConfigMap name (cortex-perses-dashboards) and type (ConfigMap) and assigns it to the desired Tilt group/label (e.g., Monitoring) using the new_name and objects parameters so Tilt will display it as a resource.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@Tiltfile`:
- Around line 291-299: The Perses helm_resource ('cortex-perses') may start
before Prometheus and its Service (prometheus-operated:9090) is available;
update the helm_resource call for 'cortex-perses' to include 'cortex-prometheus'
in the resource_deps list so Tilt waits for the Prometheus resource before
creating Perses, i.e., add 'cortex-prometheus' alongside the existing 'perses'
entry in resource_deps of the helm_resource invocation.
- Around line 300-306: The generated ConfigMap created via
watch_file('./tools/perses/dashboards') + k8s_yaml(...) is correct but won’t
show up in Tilt’s UI; add a k8s_resource declaration to surface it in the Tilt
resource list. Add a k8s_resource call that references the ConfigMap name
(cortex-perses-dashboards) and type (ConfigMap) and assigns it to the desired
Tilt group/label (e.g., Monitoring) using the new_name and objects parameters so
Tilt will display it as a resource.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: f95f27ff-7047-43d1-9be6-8696e1a60c99
📒 Files selected for processing (3)
Tiltfiletools/perses/dashboards/project.jsontools/perses/values.yaml
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
tools/perses/dashboards/cortex-compute-kvm-overview.json (1)
376-389: Use the total-capacity metric for the host count card.This panel is the odd one out: it counts hosts from a
type="utilized"series, while the capacity totals come fromcortex_kvm_host_capacity_total. Switching the count to the total metric keeps the card stable even when a host has no current workload sample.♻️ Suggested query
- "query": "count(max(cortex_kvm_host_capacity_usage{resource=\"cpu\", type=\"utilized\", availability_zone=~\"$availability_zone\", workload_type=~\"$workload_type\", cpu_architecture=~\"$cpu_architecture\", building_block=~\"$building_block\", compute_host=~\"$compute_host\", enabled=~\"$enabled\", maintenance=~\"$maintenance\", decommissioned=~\"$decommissioned\", external_customer=~\"$external_customer\"}) by (compute_host))", + "query": "count(max(cortex_kvm_host_capacity_total{resource=\"cpu\", availability_zone=~\"$availability_zone\", workload_type=~\"$workload_type\", cpu_architecture=~\"$cpu_architecture\", building_block=~\"$building_block\", compute_host=~\"$compute_host\", enabled=~\"$enabled\", maintenance=~\"$maintenance\", decommissioned=~\"$decommissioned\", external_customer=~\"$external_customer\"}) by (compute_host))",🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@tools/perses/dashboards/cortex-compute-kvm-overview.json` around lines 376 - 389, The host-count query in the TimeSeriesQuery (PrometheusTimeSeriesQuery spec under "queries") uses cortex_kvm_host_capacity_usage with type="utilized", which undercounts hosts without current utilization; replace that metric with cortex_kvm_host_capacity_total while keeping the same label filters (availability_zone, workload_type, cpu_architecture, building_block, compute_host, enabled, maintenance, decommissioned, external_customer) and the same aggregation (count(max(...) by (compute_host))) and seriesNameFormat ("Number of Hosts") so the card counts hosts based on total capacity instead of utilization.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@tools/perses/dashboards/cortex-compute-kvm-overview.json`:
- Around line 661-667: The compute_host PrometheusLabelValuesVariable is only
scoped by building_block; update its spec.matchers to mirror the upstream
filters (AZ, workload, architecture) so the host dropdown respects the same
slice. Locate the variable with labelName "compute_host" and replace the matcher
"cortex_kvm_host_capacity_total{building_block=~\"$building_block\"}" with a
matcher that includes the additional label filters, e.g.
"cortex_kvm_host_capacity_total{building_block=~\"$building_block\",
availability_zone=~\"$availability_zone\", workload=~\"$workload\",
architecture=~\"$architecture\"}" (use the exact upstream variable names used
elsewhere if they differ).
- Line 13: The dashboard JSON has a project slug mismatch: update the
metadata.project value (the "project" key in the JSON object) from
"workload-management" to the correct manifest name "workload_management" so the
dashboard will attach to the Perses project; locate the "project" entry
(currently "workload-management") and replace the hyphen with an underscore to
become "workload_management".
---
Nitpick comments:
In `@tools/perses/dashboards/cortex-compute-kvm-overview.json`:
- Around line 376-389: The host-count query in the TimeSeriesQuery
(PrometheusTimeSeriesQuery spec under "queries") uses
cortex_kvm_host_capacity_usage with type="utilized", which undercounts hosts
without current utilization; replace that metric with
cortex_kvm_host_capacity_total while keeping the same label filters
(availability_zone, workload_type, cpu_architecture, building_block,
compute_host, enabled, maintenance, decommissioned, external_customer) and the
same aggregation (count(max(...) by (compute_host))) and seriesNameFormat
("Number of Hosts") so the card counts hosts based on total capacity instead of
utilization.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 4fa6148b-50b2-4f2a-a456-cccc6596ca6a
📒 Files selected for processing (1)
tools/perses/dashboards/cortex-compute-kvm-overview.json
| "infrastructure", | ||
| "kvm" | ||
| ], | ||
| "project": "workload-management" |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
set -euo pipefail
python - <<'PY'
import json
from pathlib import Path
dashboard = json.loads(Path("tools/perses/dashboards/cortex-compute-kvm-overview.json").read_text())
project = json.loads(Path("tools/perses/dashboards/project.json").read_text())
dashboard_project = dashboard["metadata"].get("project")
project_name = project["metadata"].get("name")
print(f"dashboard.metadata.project = {dashboard_project}")
print(f"project.metadata.name = {project_name}")
if dashboard_project != project_name:
raise SystemExit("Mismatch detected")
PYRepository: cobaltcore-dev/cortex
Length of output: 196
Fix project slug mismatch: use workload_management instead of workload-management.
The dashboard's metadata.project value (workload-management) does not match the project manifest name (workload_management). This underscore-dash drift will prevent the dashboard from attaching to the Perses project.
Change line 13 from:
"project": "workload-management"to:
"project": "workload_management"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@tools/perses/dashboards/cortex-compute-kvm-overview.json` at line 13, The
dashboard JSON has a project slug mismatch: update the metadata.project value
(the "project" key in the JSON object) from "workload-management" to the correct
manifest name "workload_management" so the dashboard will attach to the Perses
project; locate the "project" entry (currently "workload-management") and
replace the hyphen with an underscore to become "workload_management".
Comment on lines
+661
to
+667
| "plugin": { | ||
| "kind": "PrometheusLabelValuesVariable", | ||
| "spec": { | ||
| "labelName": "compute_host", | ||
| "matchers": [ | ||
| "cortex_kvm_host_capacity_total{building_block=~\"$building_block\"}" | ||
| ] |
There was a problem hiding this comment.
Scope the host picker with the upstream filters.
Right now compute_host only depends on building_block, so after narrowing by AZ/workload/architecture you can still select hosts outside that slice and end up with empty panels. Mirroring the already-defined filters here keeps the dropdown coherent.
♻️ Suggested matcher
- "cortex_kvm_host_capacity_total{building_block=~\"$building_block\"}"
+ "cortex_kvm_host_capacity_total{availability_zone=~\"$availability_zone\", workload_type=~\"$workload_type\", cpu_architecture=~\"$cpu_architecture\", building_block=~\"$building_block\"}"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@tools/perses/dashboards/cortex-compute-kvm-overview.json` around lines 661 -
667, The compute_host PrometheusLabelValuesVariable is only scoped by
building_block; update its spec.matchers to mirror the upstream filters (AZ,
workload, architecture) so the host dropdown respects the same slice. Locate the
variable with labelName "compute_host" and replace the matcher
"cortex_kvm_host_capacity_total{building_block=~\"$building_block\"}" with a
matcher that includes the additional label filters, e.g.
"cortex_kvm_host_capacity_total{building_block=~\"$building_block\",
availability_zone=~\"$availability_zone\", workload=~\"$workload\",
architecture=~\"$architecture\"}" (use the exact upstream variable names used
elsewhere if they differ).
Contributor
Test Coverage ReportTest Coverage 📊: 68.6% |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Since we want to replace plutono with perses in the future here is a PR to add the perses dependency.