a tiny job in your own CI compares your Terraform state with reality and pings Slack the moment something changes outside your code. your credentials never leave your pipeline.
drift is invisible until a confusing terraform plan eats forty minutes of someone's afternoon. driftd makes it loud at exactly the right moment — and silent every other day.
the agent runs inside your own CI — the place that already has cloud access for terraform plan. only structured findings cross the wire. never state files, never secrets.
a single binary checks for drift on every pipeline run, using credentials that never leave your environment.
resource names and diffs are posted outbound to our API, like a webhook. we never connect into your infrastructure.
a precise Slack alert names the resource and the change — plus run history and a weekly digest in the dashboard.
# add five lines to any pipeline - uses: driftd/check@v1 with: token: ${{ secrets.DRIFTD_TOKEN }} dir: ./infra
the thing that touches your infrastructure is open source. audit it, fork it, pin the version.
only findings cross the boundary. there's a free mode that prints locally and sends nothing at all.
fixes arrive as pull requests a person merges. driftd never changes your infrastructure for you.
anything that runs once and prints is free. memory, aggregation and notifications are paid.
if enough people want this, we build it. that's the whole plan — and you'd be helping shape it.