Reliability Engineering

What changed five minutes before this incident,
across every system you own

Nova Rewind ingests every change signal across your stack. Deploys, feature flags, config edits, secret rotations, IAM changes, schema migrations. When an incident fires, Rewind shows you the change cone in the last hour ranked by likelihood. The "what changed?" question stops being a 20-minute war-room exercise.

Get Started Talk to Sales
app.novaaiops.com / nova-rewind
● LIVE
500+
Change source integrations
< 5s
Correlation latency
Auto
Suspect ranking
1-click
Rollback (where supported)
Every Change Source, One Stream

Deploys, flags, configs, schema, IAM, secrets

Rewind ingests change signals from CI/CD (GitHub Actions, Jenkins, Buildkite, Argo), feature-flag platforms (LaunchDarkly, Unleash, Optimizely), config tools (Consul, Terraform, Pulumi), schema migration tools (Liquibase, Flyway, sqitch), IAM (AWS IAM, GCP IAM, Okta), and secret managers (Vault, AWS Secrets, Azure Key Vault). One unified stream, one timeline.

  • ✓
    500+ integrations: every CI provider, every flag provider, every config tool ships a webhook to Rewind out of the box
  • ✓
    Custom sources via webhook: an in-house migration tool? Send POSTs to /api/rewind/event and they show up in the same timeline
  • ✓
    Service-tagged events: every change is tagged with the affected service so the cone you see is the cone that matters
app.novaaiops.com / nova-rewind · sources
The Change Cone

Suspect-ranked, not chronologically dumped

A flat timeline of every change is not useful at 3am. The change cone ranks each change by likelihood it caused the incident, using service-graph proximity, ownership, and historical incident-cause patterns. The top three changes get a confidence score and a "why we suspect this" explanation.

  • ✓
    Service-graph proximity: a deploy on the affected service is suspect 1, a deploy on an unrelated service is filtered out
  • ✓
    Historical priors: if 80% of past incidents on this service were caused by deploys, deploys get higher weight
  • ✓
    Confidence + explanation: each suspect shows a 0–100 score and a sentence on why it is suspect, no black boxes
app.novaaiops.com / nova-rewind · cone
One-Click Rollback

Where the source supports it, undo from Rewind

Many change sources support rollback (Argo, GitHub Actions, LaunchDarkly, Unleash). For those, Rewind exposes a one-click rollback button that calls the source's rollback API and writes the action to Agent Ledger. For sources that do not (raw kubectl, hand-edited configs), Rewind generates the rollback runbook for an engineer to run.

  • ✓
    Native rollback: Argo, LaunchDarkly, Optimizely, Unleash have one-click rollback wired in
  • ✓
    Generated runbook: for sources without an API, Rewind drafts the rollback steps and posts them to the war room
  • ✓
    Logged like an agent action: a rollback from Rewind writes to Agent Ledger and Approval Queue if the action class requires it
app.novaaiops.com / nova-rewind · rollback
Drift Detection

Catch silent changes too

Not every change comes through a CI/CD pipeline. Engineers ssh and edit a config. A cloud provider auto-rotates a key. Rewind detects state drift across cloud resources, Kubernetes manifests, and config files, and adds those drifts to the change cone alongside the recorded events. Anything that mutates is surfaced.

  • ✓
    Cloud drift: detected by reconciling AWS/GCP/Azure resource state against the last-known IaC plan, every 5 minutes
  • ✓
    Kubernetes drift: kubectl edits show up next to your normal deploys with the diff, the user, and the timestamp
  • ✓
    Config-file drift: agents on each box watch /etc and report unsigned edits to Rewind, scoped to the service
app.novaaiops.com / nova-rewind · drift
Video walkthrough coming soon

Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.

Stop guessing what changed

Rewind is the answer to "did anything ship today?" already on your screen by the time the page hits.

Get Started Request a Demo