Fix landscaping-assistant dashboard metric names #403
No reviewers
Labels
No labels
domain:backend
domain:devops
domain:frontend
status:approved
status:in-progress
status:needs-fix
status:qa
type:bug
type:devops
type:feature
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
ldraney/pal-e-platform!403
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "fix-dashboard-metric-names"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Closes #401
Summary
Dashboard queries used assumed metric names from docs. Validated against live
/metricsendpoint and corrected all 8 mismatched names.Changes
terraform/dashboards/landscaping-assistant-golden-signals.json— fix all PromQL expressions to match actual yabeda outputTest Plan
curl https://landscaping-assistant.tail5b443a.ts.net/metricstofu applyto update ConfigMapReview Checklist
_seconds_suffix (yabeda convention)threads_prefix)Acceptance Criteria
Related Notes
/metricsendpointPR #403 Review
DOMAIN REVIEW
Tech stack: Terraform/Grafana dashboard JSON (infrastructure/observability domain).
File changed:
terraform/dashboards/landscaping-assistant-golden-signals.json(1 file)The diff is 659 additions / 193 deletions, but the vast majority is
json.dumpsreformatting (compact single-line objects expanded to multi-line). The actual semantic changes are 8 metric name corrections and 3 panel title/legend updates. I verified each substantive change:Histogram names -- adding
_seconds_suffix (yabeda convention):rails_request_duration_bucket->rails_request_duration_seconds_bucket(p50, p95, p99 panels) -- CORRECTrails_db_runtime_bucket->rails_db_runtime_seconds_bucket-- CORRECTrails_view_runtime_bucket->rails_view_runtime_seconds_bucket-- CORRECTPuma gauge names -- removing
threads_prefix:puma_threads_running->puma_running(2 occurrences incl. utilization calc) -- CORRECTpuma_threads_total->puma_max_threads(2 occurrences incl. utilization calc) -- CORRECTpuma_threads_backlog->puma_backlog-- CORRECTpuma_threads_pool_capacity->puma_pool_capacity-- CORRECTNon-existent metric replacement:
puma_workers->puma_requests_count-- CORRECT (panel title and legend also updated)Utilization formula consistency check: The utilization percentage formula
puma_running / clamp_min(puma_max_threads, 1) * 100correctly uses the updated metric names on both sides of the division. No stale references remain.Panel metadata updates: Title "Puma Workers & Pool Capacity" -> "Puma Requests & Pool Capacity" and legend labels updated to match new metric semantics ("max threads", "requests served"). These are accurate.
No structural changes: Panel IDs, grid positions, datasource references, schema version, and template variables are unchanged. The reformatting is cosmetic only.
BLOCKERS
None.
NITS
JSON formatting churn: The reformatting from compact to expanded JSON creates noise in the diff (659+/193- for ~8 actual changes). Future dashboard edits could benefit from a consistent formatter (e.g.,
python -m json.toolor a pre-commit hook) so formatting is always normalized and diffs only show real changes. Not blocking -- this is a one-time cleanup.Test plan has unchecked items: The PR body shows
tofu applyand "Panels show data in Grafana" as unchecked. These are post-merge validation steps, which is fine, but worth confirming they will be done after merge.SOP COMPLIANCE
fix-dashboard-metric-names, not401-fix-dashboard-metric-names. Missing the issue number prefix per{issue-number}-{kebab-case-purpose}convention. Non-blocking since the parent issue #401 is already closed and this is a hotfix follow-up to PR #402.PROCESS OBSERVATIONS
/metricsendpoint, which is the right approach._seconds_histogram naming convention so future dashboard authors don't repeat the same mistake.VERDICT: APPROVED
Clean, well-scoped fix. All 8 metric name corrections are consistent with yabeda/puma-stats-exporter conventions. No stale references, formula consistency verified, panel metadata updated to match. The JSON reformatting noise is cosmetic and harmless.
View command line instructions
Checkout
From your project repository, check out a new branch and test the changes.