rootecho

Catch recurring root causes across postmortems. Heavyweight incident platforms (rootly, incident.io) flag when a new incident shares a root cause with a past one — teams without that budget just... don't find out, until the same failure bites twice. rootecho does the comparison locally: no account, no server, your incident history lives in your repo.

pip install rootecho
rootecho init incident.json      # scaffold a postmortem
rootecho add incident.json       # record it, flag any echo of a past root cause
rootecho check incident.json     # same check, no recording — use as a CI gate

This is the Python build — a Node build (npx rootecho) exists too and reads the exact same .rootecho/history.jsonl, so a team split across both ecosystems shares one history.

Why

"We use rootly to track this automatically. It flags when incidents have the same root cause as previous ones."

That's a paid, hosted feature. For everyone else, a postmortem gets written, action items get filed, and six months later the exact same root cause causes the exact same outage — because nobody had a system for "hey, we've seen this before, and last time's fix never shipped." rootecho is that system, as a zero-dependency local CLI.

How it works

1. Each postmortem is one JSON record — root_cause (free text) and/or root_cause_tags (curated labels), plus action_items with a status.
2. add/check compare it against your history using Jaccard similarity over tags (primary signal) blended with free-text overlap (secondary). No ML dependency, no network call.
3. A match above the threshold prints the past incident's action items — so you see immediately whether last time's fix ever actually shipped.

Incident format

{
  "id": "INC-2026-014",
  "date": "2026-07-03",
  "title": "Payment webhook retries exhausted",
  "root_cause": "webhook retry queue misconfigured to drop after 3 attempts, no dead-letter fallback",
  "root_cause_tags": ["webhook", "retry-queue", "dead-letter", "config"],
  "action_items": [
    { "id": "AI-1", "description": "Add dead-letter queue for webhook retries", "owner": "alice", "status": "open", "due_date": "2026-07-20" }
  ]
}

Only id and one of root_cause/root_cause_tags are required. rootecho init writes a starter file.

Example

$ rootecho add inc-2026-014.json
⚠ root cause echo detected for "INC-2026-014":

  INC-2026-003 (2026-03-15) — 100% similar root cause
  Payment webhook retries exhausted
    ✓ Add retry backoff [done]
    ✗ Add monitoring alert for queue depth [open] — 93d overdue
  → 1 action item(s) from this past incident were never finished.

recorded to .rootecho/history.jsonl

Commands

rootecho add <file>      # record + compare (always exits 0 on success)
rootecho check <file>    # compare only, no recording — exit 1 if an echo is found
rootecho list [--json]   # show recorded incidents and open action-item counts
rootecho init [file]     # write a starter incident.json

Flags: --dir <path> (state location), --threshold <0-1> (default 0.34, lower = more sensitive), --json, --force (init: overwrite; add: allow a duplicate id).

Storage

History is a JSON-Lines file at .rootecho/history.jsonl, local to your project by default (not your home directory) — the idea is your team commits it alongside the postmortems it describes, so git blame/git log on the file doubles as an incident timeline. Override the location with --dir or $ROOTECHO_HOME.

Exit codes

License

MIT