BlameTrail
Deploy Tracking

Overview

How BlameTrail tracks deploys, enriches them with GitHub metadata, and uses temporal proximity to identify suspect code changes when incidents occur.

Deploy tracking is a core part of how BlameTrail connects incidents to the code changes that caused them. Every time your team deploys, BlameTrail records the event, enriches it with context from GitHub, and holds it ready for suspect scoring the moment something goes wrong.

How it works

The deploy tracking pipeline has four stages:

  1. Ingest — Your CI/CD pipeline sends a deploy event to BlameTrail via a webhook. The event includes the commit SHA, branch, deployer, environment, and an optional list of commits.
  2. Link — BlameTrail associates the deploy with the correct service based on the webhook token. Each token is scoped to a single service and tenant.
  3. Enrich — If the service is linked to a GitHub repository and a GitHub token is configured, BlameTrail fetches additional metadata in the background: full commit messages, linked pull request titles, authors, and changed files ranked by relevance.
  4. Score — When an incident opens, BlameTrail examines all deploys to the affected service from the last 60 minutes and ranks them by time proximity to the first failure. The result is a prioritized list of suspect deploys with full context.

What gets recorded

Each deploy event captures:

FieldRequiredDescription
commit_shaYesThe Git commit SHA that was deployed
commit_messageNoThe head commit message
branchNoThe branch that was deployed
deployed_byNoWho triggered the deploy
environmentNoTarget environment (e.g., production, staging)
versionNoA version string or tag
descriptionNoFree-text description of the deploy
urlNoLink to the deploy in your CI/CD system
commitsNoArray of commits included in the deploy

Why it matters

Without deploy data, incident investigation starts with a blank page. With it, BlameTrail can answer the question your team asks first: "What changed?"

  • Suspect scoring immediately surfaces the most likely cause when an incident opens, cutting triage time from minutes to seconds.
  • Commit enrichment adds PR context and changed file lists so you can evaluate suspects without leaving BlameTrail.
  • AI summaries reference suspect deploys and their changes, providing a narrative explanation of what likely went wrong.
  • Timeline view shows deploys alongside incidents on a unified timeline, making it easy to spot patterns.

Next steps

On this page