Dev Health Panels

Opinionated Grafana panel plugin for teaching developer health through visual signals (not scores).

Panels

• Developer Landscape: quadrant scatter for normalized tradeoffs (0.5 / 0.5 boundaries).
• Hotspot Explorer: table with sparkline trends and donut drivers per file.
• Investment Flow: Sankey-style flow of effort from investment areas to outcomes.

Development

npm install
npm run dev

Build

npm run build

Unsigned plugin

This plugin loads unsigned. Ensure Grafana has GF_PLUGINS_ALLOW_LOADING_UNSIGNED_PLUGINS=devhealth-ops-panel.

Dashboard Configuration

The following queries and variables are recommended for use with the standard dev-health-ops dashboards.

Panel Query: Set Format to Table in the query options.

SELECT
  max(v.as_of_day) as as_of_day,
  map_name,
  team_id,
  identity_id,
  argMax(x_raw, v.as_of_day) as x_raw,
  argMax(y_raw, v.as_of_day) as y_raw,
  argMax(x_norm, v.as_of_day) as x_norm,
  argMax(y_norm, v.as_of_day) as y_norm
FROM stats.v_ic_landscape_points AS v
WHERE map_name = 'churn_throughput' -- REPLACE with 'cycle_throughput' or 'wip_throughput'
  AND v.as_of_day >= toDate(toDateTime(intDiv($__from, 1000)))
  AND v.as_of_day < toDate(toDateTime(intDiv($__to, 1000)))
  AND team_id IN (${team_id:sqlstring})
GROUP BY map_name, team_id, identity_id
ORDER BY team_id, identity_id

Hotspot Explorer (table + sparklines)

Variables:

• repo_id (Query): SELECT repo AS __text, repo AS __value FROM repos ORDER BY repo (Multi + IncludeAll + Regex)

Panel Query A (Facts):

SELECT
    metrics.repo_id AS repo_id,
    metrics.path AS file_path,
    sum(metrics.churn) AS churn_loc_30d, -- Corrected Alias
    lookup.cyclomatic_total AS cyclomatic_total,
    lookup.ownership_concentration AS ownership_concentration,
    0 AS incident_count,
    log1p(churn_loc_30d) AS churn_signal,
    log1p(coalesce(cyclomatic_total, 0)) AS complexity_signal,
    coalesce(ownership_concentration, 0) AS ownership_signal,
    log1p(incident_count) AS incident_signal,
    (0.5 * churn_signal + 0.3 * complexity_signal + 0.2 * ownership_signal) AS risk_score
FROM stats.file_metrics_daily AS metrics
INNER JOIN repos AS r ON r.id = metrics.repo_id
LEFT JOIN (
    SELECT
        repo_id,
        file_path,
        argMax(cyclomatic_total, computed_at) AS cyclomatic_total,
        argMax(blame_concentration, computed_at) AS ownership_concentration
    FROM stats.file_hotspot_daily
    GROUP BY repo_id, file_path
) AS lookup
    ON lookup.repo_id = metrics.repo_id
    AND lookup.file_path = metrics.path
WHERE match(r.repo, '${repo_id:regex}')
  AND metrics.day >= toDate(toDateTime(intDiv($__from, 1000)))
  AND metrics.day < toDate(toDateTime(intDiv($__to, 1000)))
GROUP BY metrics.repo_id, metrics.path, lookup.cyclomatic_total, lookup.ownership_concentration
ORDER BY risk_score DESC
LIMIT 50

Panel Query B (Sparklines):

SELECT
    metrics.path AS file_path,
    metrics.day,
    sum(metrics.churn) AS churn_loc
FROM stats.file_metrics_daily AS metrics
INNER JOIN repos AS r ON r.id = metrics.repo_id
WHERE match(r.repo, '${repo_id:regex}')
  AND metrics.day >= toDate(toDateTime(intDiv($__from, 1000)))
  AND metrics.day < toDate(toDateTime(intDiv($__to, 1000)))
GROUP BY metrics.path, metrics.day
ORDER BY file_path, day

Investment Flow (Sankey)

Variables:

• team_id (Query): SELECT DISTINCT team_id AS __text, team_id AS __value FROM (SELECT team_id FROM stats.team_metrics_daily UNION ALL SELECT team_id FROM stats.investment_metrics_daily) ORDER BY team_id (Multi + IncludeAll)

Panel Query:

SELECT
    source,
    target,
    sum(delivery_units) AS value
FROM stats.v_investment_flow_edges
WHERE team_id IN (${team_id:sqlstring})
  AND day >= toDate(toDateTime(intDiv($__from, 1000)))
  AND day < toDate(toDateTime(intDiv($__to, 1000)))
GROUP BY source, target
HAVING value > 0
ORDER BY value DESC