DORA - Median Lead Time for Changes

What is this metric?

The median amount of time for a code change to be deployed into production.

Why is it important?

This metric measures the time it takes to a code change to the production environment and reflects the speed of software delivery. A lower average change preparation time means that your team is efficient at coding and deploying your project.

Which dashboard(s) does it exist in

DORA dashboard. See live demo.

How is it calculated?

This metric is quite similar to PR Cycle Time. The difference is that 'Lead Time for Changes' uses a different method to filter PRs.

1. Find the PRs' associated deployment commits whose finished_date falls into the time range that users select.
2. Find the associated pull requests of the commits diff between two consecutive successful deployment commits in the production environment.
3. Calculate the PRs' median cycle time. This will be the Median Lead Time for Changes.

PR cycle time is pre-calculated by the dora plugin during every data collection. You can find it in pr_cycle_time in table.project_pr_metrics of DevLake's database.

Below are the benchmarks for different development teams from Google's report. However, it's difficult to tell which group a team falls into when the team's median lead time for changes is between one week and one month. Therefore, DevLake provides its own benchmarks to address this problem:

Groups	Benchmarks	DevLake Benchmarks
Elite performers	Less than one hour	Less than one hour
High performers	Between one day and one week	Less than one week
Medium performers	Between one month and six months	Between one week and six months
Low performers	More than six months	More than six months

Source: 2021 Accelerate State of DevOps, Google

Data Sources Required

• Deployments from Jenkins, GitLab CI, GitHub Action, BitBucket Pipelines, Webhook, etc.
• Pull Requests from GitHub PRs, GitLab MRs, BitBucket PRs, Azure DevOps PRs, etc.

Transformation Rules Required

Define deployment in data transformations while configuring the blueprint of a project to let DevLake know what CI records can be regarded as deployments.

SQL Queries

If you want to measure the monthly trend of median lead time for changes as the picture shown below, run the following SQL in Grafana.

with _pr_stats as (
-- get the cycle time of PRs deployed by the deployments finished each month
    SELECT
        distinct pr.id,
        date_format(cdc.finished_date,'%y/%m') as month,
        ppm.pr_cycle_time
    FROM
        pull_requests pr
        join project_pr_metrics ppm on ppm.id = pr.id
        join project_mapping pm on pr.base_repo_id = pm.row_id
        join cicd_deployment_commits cdc on ppm.deployment_commit_id = cdc.id
    WHERE
        pm.project_name in ($project) 
        and pr.merged_date is not null
        and ppm.pr_cycle_time is not null
        and $__timeFilter(cdc.finished_date)
),

_find_median_clt_each_month_ranks as(
    SELECT *, percent_rank() over(PARTITION BY month order by pr_cycle_time) as ranks
    FROM _pr_stats
),

_clt as(
    SELECT month, max(pr_cycle_time) as median_change_lead_time
    FROM _find_median_clt_each_month_ranks
    WHERE ranks <= 0.5
    group by month
),

_calendar_months as(
-- to   deal with the month with no incidents
    SELECT date_format(CAST((SYSDATE()-INTERVAL (month_index) MONTH) AS date), '%y/%m') as month
    FROM ( SELECT 0 month_index
            UNION ALL SELECT   1  UNION ALL SELECT   2 UNION ALL SELECT   3
            UNION ALL SELECT   4  UNION ALL SELECT   5 UNION ALL SELECT   6
            UNION ALL SELECT   7  UNION ALL SELECT   8 UNION ALL SELECT   9
            UNION ALL SELECT   10 UNION ALL SELECT  11
        ) month_index
    WHERE (SYSDATE()-INTERVAL (month_index) MONTH) > SYSDATE()-INTERVAL 6 MONTH 
)

SELECT 
    cm.month,
    case 
        when _clt.median_change_lead_time is null then 0 
        else _clt.median_change_lead_time/60 end as median_change_lead_time_in_hour
FROM 
    _calendar_months cm
    left join _clt on cm.month = _clt.month
ORDER BY 1

If you want to measure in which category your team falls as in the picture shown below, run the following SQL in Grafana.

with _pr_stats as (
-- get the cycle time of PRs deployed by the deployments finished in the selected period
    SELECT
        distinct pr.id,
        ppm.pr_cycle_time
    FROM
        pull_requests pr 
        join project_pr_metrics ppm on ppm.id = pr.id
        join project_mapping pm on pr.base_repo_id = pm.row_id
        join cicd_deployment_commits cdc on ppm.deployment_commit_id = cdc.id
    WHERE
      pm.project_name in ($project) 
        and pr.merged_date is not null
        and ppm.pr_cycle_time is not null
        and $__timeFilter(cdc.finished_date)
),

_median_change_lead_time_ranks as(
    SELECT *, percent_rank() over(order by pr_cycle_time) as ranks
    FROM _pr_stats
),

_median_change_lead_time as(
-- use median PR cycle time as the median change lead time
    SELECT max(pr_cycle_time) as median_change_lead_time
    FROM _median_change_lead_time_ranks
    WHERE ranks <= 0.5
)

SELECT 
  CASE
    WHEN median_change_lead_time < 60 then "Less than one hour"
    WHEN median_change_lead_time < 7 * 24 * 60 then "Less than one week"
    WHEN median_change_lead_time < 180 * 24 * 60 then "Between one week and six months"
    WHEN median_change_lead_time >= 180 * 24 * 60 then "More than six months"
    ELSE "N/A.Please check if you have collected deployments/incidents."
    END as median_change_lead_time
FROM _median_change_lead_time

How to improve?

• Break requirements into smaller, more manageable deliverables
• Optimize the code review process
• "Shift left", start QA early and introduce more automated tests
• Integrate CI/CD tools to automate the deployment process