Why Every DevOps Engineer Should Learn Power BI in 2026

STAR SCHEMA FOR A DEVOPS DASHBOARD

Fact tables
  Deployments   : DeployId, Service, Environment, DeployTime,
                  FirstCommitTime, Status, CausedIncident (TRUE/FALSE)
  Incidents     : IncidentId, Service, Severity, StartTime,
                  ResolvedTime, LinkedDeployId
  PipelineRuns  : RunId, Pipeline, Result, StartTime, DurationSec

Dimension tables
  Date          : Date, Week, Month, Quarter (mark as date table)
  Service       : Service, Team, Tier

Relationships (one-to-many, single direction)
  Date[Date]       -> Deployments[DeployDate]
  Date[Date]       -> Incidents[StartDate]
  Service[Service] -> Deployments[Service]
  Service[Service] -> Incidents[Service]

Date =
ADDCOLUMNS (
    CALENDAR ( DATE ( 2026, 1, 1 ), DATE ( 2026, 12, 31 ) ),
    "Year", YEAR ( [Date] ),
    "Month", FORMAT ( [Date], "YYYY-MM" ),
    "Week Start", [Date] - WEEKDAY ( [Date], 2 ) + 1,
    "Quarter", "Q" & QUARTER ( [Date] )
)

// Then: Table tools > Mark as date table > choose [Date]
// Add calculated columns in fact tables:
// DeployDate = DATEVALUE ( Deployments[DeployTime] )

// 1) Azure DevOps pipeline runs via OData Feed
// Get Data > OData Feed, then paste a URL like:
// https://analytics.dev.azure.com/{org}/{project}/_odata/v4.0-preview/PipelineRuns?$filter=CompletedDate ge 2026-08-01T00:00:00Z&$select=PipelineRunId,RunNumber,RunOutcome,StartedDate,CompletedDate,RunDurationSeconds
// (Check the Analytics docs for the current version string.)

// 2) GitHub Actions runs for a public repo (Blank Query)
let
    Source = Json.Document(
        Web.Contents("https://api.github.com",
            [RelativePath = "repos/OWNER/REPO/actions/runs",
             Query = [per_page = "100", status = "completed"],
             Headers = [Accept = "application/vnd.github+json"]])),
    Runs = Table.FromList(Source[workflow_runs], Splitter.SplitByNothing()),
    Expanded = Table.ExpandRecordColumn(Runs, "Column1",
        {"id", "name", "conclusion", "head_branch", "created_at", "updated_at"}),
    Typed = Table.TransformColumnTypes(Expanded,
        {{"created_at", type datetimezone}, {"updated_at", type datetimezone}})
in
    Typed

# pip install requests pandas
# Exports resolved incidents from Jira Cloud to CSV for Power BI
import os, requests, pandas as pd

base = "https://your-site.atlassian.net"
auth = (os.environ["JIRA_EMAIL"], os.environ["JIRA_API_TOKEN"])
params = {
    "jql": "project = OPS AND issuetype = Incident AND resolved >= -90d",
    "fields": "created,resolutiondate,priority,components",
    "maxResults": 100,
}

rows, token = [], None
while True:
    if token:
        params["nextPageToken"] = token
    r = requests.get(f"{base}/rest/api/3/search/jql", params=params, auth=auth)
    r.raise_for_status()
    data = r.json()
    for i in data["issues"]:
        f = i["fields"]
        rows.append({"key": i["key"], "created": f["created"],
                     "resolved": f["resolutiondate"],
                     "priority": (f.get("priority") or {}).get("name")})
    token = data.get("nextPageToken")
    if not token:
        break

pd.DataFrame(rows).to_csv("incidents.csv", index=False)

Prod Deploys =
CALCULATE (
    COUNTROWS ( Deployments ),
    Deployments[Environment] = "production",
    Deployments[Status] = "succeeded"
)

Deploys per Week =
DIVIDE ( [Prod Deploys], DISTINCTCOUNT ( 'Date'[Week Start] ) )

Lead Time (median hrs) =
MEDIANX (
    FILTER ( Deployments, Deployments[Environment] = "production" ),
    DATEDIFF ( Deployments[FirstCommitTime], Deployments[DeployTime], HOUR )
)

Change Failure Rate =
DIVIDE (
    CALCULATE ( [Prod Deploys], Deployments[CausedIncident] = TRUE () ),
    [Prod Deploys]
)

MTTR (min) =
AVERAGEX (
    FILTER ( Incidents, NOT ISBLANK ( Incidents[ResolvedTime] ) ),
    DATEDIFF ( Incidents[StartTime], Incidents[ResolvedTime], MINUTE )
)

MTTR Sev1 (min) =
CALCULATE ( [MTTR (min)], Incidents[Severity] = "Sev1" )

Open Incidents =
CALCULATE ( COUNTROWS ( Incidents ), ISBLANK ( Incidents[ResolvedTime] ) )

MTTR vs Last Month =
VAR cur = [MTTR (min)]
VAR prev = CALCULATE ( [MTTR (min)], DATEADD ( 'Date'[Date], -1, MONTH ) )
RETURN DIVIDE ( cur - prev, prev )

DEVOPS DELIVERY DASHBOARD - PAGE LAYOUT

Page 1: Overview
  Top row    : 4 KPI cards - Deploys/week, Lead time (median),
               Change failure rate, MTTR
  Middle     : Weekly deploy column chart + lead time trend line
  Bottom     : MTTR by severity (bar), open incidents (card)
  Slicers    : Date range, Team, Service, Environment

Page 2: Incidents
  Table      : Incident, service, severity, start, minutes to resolve
  Chart      : Incidents per week by service (stacked column)
  Drill-thru : From service to its incident list

Page 3: Definitions (text box - see next tab)

METRIC DEFINITIONS (put this on the report)

Deployment   : A succeeded pipeline run to the production
               environment. Re-runs of the same commit count once.
Lead time    : Hours from first commit in the release to the
               production deploy. Median shown, not average.
Failure      : A production deploy linked to a Sev1/Sev2 incident
               or rolled back within 24 hours.
Incident     : Starts at first alert or ticket creation, whichever
               is earlier; ends when service is restored.
Owner        : Platform team. Questions: #devops-metrics channel
Last reviewed: __________

REFRESH AND GOVERNANCE CHECKLIST

[ ] Data sources documented (OData URL, CSV paths, API scripts)
[ ] Export scripts scheduled before Power BI refresh time
[ ] Credentials stored in the gateway/service, not in queries
[ ] On-premises data gateway set up if sources are internal
[ ] Scheduled refresh configured within licence limits
[ ] Refresh failure email goes to a team mailbox
[ ] Row-level security if teams must see only their data
[ ] Workspace access reviewed every quarter
[ ] 'Last refreshed' timestamp visible on every page