Software Delivery
Software Delivery brings continuous integration and continuous delivery (CI/CD) visibility directly into SolarWinds Observability SaaS, alongside the applications, databases, and infrastructure you already monitor. Instead of switching between your CI/CD provider for build health and monitoring tools for production health, Software Delivery correlates both in one platform, allowing you to trace production incidents back to the specific pipeline run, commit, and pull request that introduced the change.
Software Delivery provides the following features:
-
CI visibility: Track the health, reliability, duration, and failure rates of your deployment pipelines. Quickly pinpoint which pipelines fail most frequently, which suffer from long execution times, and where engineering hours are lost to build failures.
-
Kubernetes & APM: Integrations with Kubernetes and APM show exactly when a new deployment went live, and let you trace it back to the commit or pipeline run that produced it.
To monitor Software Delivery, integrate SolarWinds Observability SaaS with your GitHub account.
How it works
Software Delivery uses an agentless integration. You do not need to install collectors or agents inside your build runners.
-
You install the dedicated SolarWinds Observability GitHub App on your GitHub organization or repositories.
-
GitHub sends webhook events (workflow_run, workflow_job, and pull request events) to SolarWinds Observability SaaS.
-
SolarWinds Observability SaaS converts these events into standard entities, traces, and metrics following OpenTelemetry CI/CD semantic conventions.
Supported providers
GitHub Actions (hosted on github.com)
GitHub Enterprise Server (on-premises) and other CI/CD providers are not currently supported.
Monitored entities
Software Delivery creates two primary entity types in SolarWinds Observability SaaS:
-
CI/CD Pipeline: Represents a named, repeating workflow in a repository.
-
CI/CD Pipeline Run: Represents a single execution of a pipeline.
For a summary of your CI/CD pipelines connected to SolarWinds Observability SaaS, click Software Delivery in the side navigation. The Software Delivery area overview shows the health of all monitored CI/CD pipelines and an overall success rate, as well as widgets that offer quick insights into failing workflows, longest-running pipelines, and time lost to build failures. An inventory of all connected pipelines and their recent execution details is available on the Pipelines tab. See Software Delivery area overview.
Metrics for pipeline entities are also available as widgets in the Entity Explorer under the Software Delivery category. You can filter, search, alert on, and add pipeline entities to custom dashboards. Individual runs are accessed through the pipeline's Runs tab or in Analyze > Traces.
A pipeline's identity is defined by its workflow name and its repository URL:
-
Renaming a workflow creates a new pipeline entity. Earlier runs remain associated with the old pipeline entity.
-
Renaming the repository creates a new pipeline entity.
-
Moving or renaming the workflow file keeps the same pipeline entity, provided the workflow's name: attribute remains unchanged. (If no name: attribute is specified, GitHub names the workflow after its file path; moving an unnamed file creates a new pipeline).
-
Two workflow files with identical names in the same repository are represented as one pipeline.
How runs are calculated across views
Different widgets in SolarWinds Observability SaaS serve different analytical purposes, which affects how run outcomes are categorized:
| Widget | Counted as failure | Canceled and skipped runs |
|---|---|---|
| Overall Success Rate | Any completed run that did not succeed. | Excluded from calculations entirely. |
| Top 10 Failing Pipelines, Failed Executions, Inspector panel | Failed, timed-out, and startup-failure runs. | Not counted as failures, but included in the total run volume denominator. |
| Time Lost to Failures, Treemap color | Failed, timed-out, and startup-failure runs. | Excluded from failure counts. |
| Runs tab Success Rate donut | Failed, timed-out, and startup-failure runs. | Counted as failures. |
| Pipeline Health Score | Any run that did not complete with success. | Excluded from calculations entirely. |
Ingested data
SolarWinds Observability SaaS ingests webhook events from your CI/CD provider and maps them to OpenTelemetry CI/CD semantic conventions. Collected data includes:
-
Pipeline name, run ID, run number, result, state, and duration.
-
Job and step names, results, execution durations, and runner identifiers.
-
Repository name, repository URL, and organization name.
-
Branch name, branch type, and commit revision hash.
-
Pull request title (for PR-triggered workflows).
-
Direct URLs back to the workflow run and individual jobs in GitHub.
Data retention
Software Delivery data is kept for the standard retention period of each data type. For more information, see Data retention period for each data type.
-
Pipeline run traces: Retained for 30 days.
-
Deployment events: Retained for 30 days.
-
Metrics: Retained at 1-minute full resolution for 7 days, 10-minute rollups for 30 days, and hourly rollups for up to 549 days.