Skip to content

DORA Metrics

DORA metrics measure the performance of your software delivery process. Originally developed by the DevOps Research and Assessment team (now part of Google Cloud), these four metrics have become the standard way to evaluate how effectively a team ships changes. Rime applies the same framework to your data pipelines, treating pipeline runs as deployments and tracking failures and recovery times automatically.

The four metrics

Deployment Frequency

How often you successfully deploy changes to production. In Rime’s context, a “deployment” is a successful pipeline run that moves data through your stack — extraction, transformation, and loading.

Higher deployment frequency generally indicates a healthier pipeline with smaller, more frequent changes that are easier to debug when something goes wrong.

Lead Time for Changes

The elapsed time from when a change is committed (e.g., a new connector configuration, a transformation model update, or a pipeline modification) to when it runs successfully in production.

Shorter lead times mean your team can respond quickly to data source changes, schema updates, or new business requirements.

Change Failure Rate

The percentage of pipeline runs that result in a failure. This includes connector sync failures, dbt run errors, infrastructure apply failures, and any pipeline step that does not complete successfully.

A low change failure rate indicates that your pipelines are stable and that changes are well-tested before they reach production.

Time to Restore Service

The elapsed time from when a failure is detected to when the pipeline is running successfully again. This measures how quickly your team identifies and resolves data pipeline issues.

Rime measures this as the time between an alert firing (or a pipeline run failing) and the next successful run of the same pipeline or connector.

How Rime collects DORA data

Rime calculates DORA metrics automatically from data it already collects:

MetricData source
Deployment FrequencyCount of successful pipeline runs per time period
Lead Time for ChangesTime between configuration change and next successful run
Change Failure RateFailed runs / total runs over the time period
Time to RestoreTime between failure detection and next successful run

No additional instrumentation is required. If you are running pipelines through Rime, the data is already being collected.

What counts as a deployment

By default, all pipeline runs count as deployments. You can customize this by selecting which pipelines should be included in DORA tracking:

  • All pipelines (default) — every pipeline run in the project is counted
  • Selected pipelines — only runs of specific pipelines count as deployments
  • Tagged pipelines — only pipelines with a specific tag (e.g., “production”) are included

This is configured in Project > Settings > DORA Metrics.

Data sources in detail

Pipeline execution history feeds Deployment Frequency and Change Failure Rate. Every pipeline run is recorded with its start time, end time, status (success/failure), and which steps succeeded or failed.

Alert resolution times feed Time to Restore. When a pipeline failure triggers an alert and the alert is later resolved (either by auto-resolution when the pipeline succeeds or by manual resolution), the duration is recorded.

Connector run success rates are included in Change Failure Rate. Connector syncs that fail contribute to the overall failure rate, even if they are not part of a formal pipeline.

The DORA dashboard

The DORA dashboard is available at Project > DORA Metrics. It displays:

Metric cards

Each of the four metrics is displayed as a card showing:

  • The current value (e.g., “4.2 runs/day” for Deployment Frequency)
  • The trend compared to the previous period (up, down, or stable)
  • The DORA classification (Elite, High, Medium, Low)

Trend charts

Each metric has a time-series chart showing its value over the selected time range. You can view data for the last 7 days, 30 days, 90 days, or a custom range. The chart includes the classification thresholds as horizontal reference lines so you can see how close you are to the next level.

Classification summary

A summary panel shows your overall DORA classification based on all four metrics. The classification is determined by the lowest-performing metric — if three metrics are Elite but one is Medium, your overall classification is Medium.

Classification thresholds

Rime uses the thresholds from the DORA research to classify performance:

Deployment Frequency

ClassificationThreshold
EliteMultiple times per day
HighOnce per day to once per week
MediumOnce per week to once per month
LowLess than once per month

Lead Time for Changes

ClassificationThreshold
EliteLess than 1 hour
High1 hour to 1 day
Medium1 day to 1 week
LowMore than 1 week

Change Failure Rate

ClassificationThreshold
Elite0-5%
High5-10%
Medium10-15%
LowMore than 15%

Time to Restore Service

ClassificationThreshold
EliteLess than 1 hour
High1 hour to 1 day
Medium1 day to 1 week
LowMore than 1 week

These thresholds are based on the original DORA research applied to software delivery. Data pipeline workloads may have different natural cadences (e.g., daily batch pipelines will never achieve “multiple times per day” deployment frequency). Interpret the classifications in the context of your pipeline design.

Setup

DORA tracking is enabled by default for all projects. To configure it:

  1. Go to Project > Settings > DORA Metrics
  2. Select which pipelines count as “deployments” (all, selected, or tagged)
  3. Optionally adjust the reporting time zone

No other setup is required. Metrics begin populating as soon as pipeline runs are recorded.

Minimum data requirements

DORA metrics require at least 7 days of pipeline run history to calculate meaningful values. Until that threshold is reached, the dashboard displays a “Collecting data” message for each metric.

Exporting DORA data

You can export DORA metric history as CSV for use in external reporting tools or management dashboards. Click Export on the DORA dashboard and select the time range and format.

The API also provides endpoints for retrieving DORA metrics programmatically, which is useful for integrating with CI/CD dashboards or executive reporting tools.

Next steps