> ## Documentation Index > Fetch the complete documentation index at: https://docs.whawit.ai/llms.txt > Use this file to discover all available pages before exploring further. # Backing up Opsgenie before it shuts down > A free, open-source command that copies an entire Opsgenie account to plain JSON — configuration and the full alert history — before Atlassian deletes it on 5 April 2027 Atlassian deletes every unmigrated Opsgenie account on **5 April 2027**. The built-in export is an iCalendar file of the next twelve months of schedules; it does not carry routing rules, escalation repeat settings, overrides, alert and notification policies, heartbeats, user contact methods, integrations, or a single alert of history. `opsgenie-export` is a command-line tool WHAWIT publishes under the MIT license. It reads the whole account through the REST API and writes plain JSON files to a folder on your machine. It is useful whether or not you move to WHAWIT: the [importer](/on-call/opsgenie) recreates schedules and escalation policies; this tool keeps everything else, including the history the importer deliberately leaves behind. ```bash theme={null} npx @usewhawit/opsgenie-export --api-key ``` The tool is **read-only**. It talks to `api.opsgenie.com` (or the EU host) and nothing else — no sign-up, no telemetry, no upload. Nothing in Opsgenie is modified, and the folder never leaves your disk unless you move it. ## What you need `npx` downloads and runs the current version; nothing is installed globally. **Settings → API key management → Add new API key**, with **Read** and **Configuration Access**. A team integration key only sees alerts and fails with HTTP 403 on the configuration endpoints. The same key works for the [importer](/on-call/opsgenie). Add `--region eu` for accounts on `api.eu.opsgenie.com`. ## What you get ``` opsgenie-export-20270105-093000Z/ manifest.json counts per section, warnings, what was redacted account.json users.json every user, with contact methods and notification rules teams.json every team, with members, routing rules and team roles schedules.json every schedule, rotations expanded, with all overrides on-calls.json who was on call in each schedule at export time escalations.json rules, delays and repeat settings integrations.json configuration per integration (credentials redacted by default) policies.json alert and notification policies, global and per team maintenance.json heartbeats.json services.json forwarding-rules.json roles.json alerts/2026-03.ndjson one file per month, one alert per line, oldest first incidents/2026-03.ndjson state.json completed months, so --resume can pick up where a run stopped ``` History is walked month by month, sorted by creation time. Opsgenie refuses any search past 20,000 results, so a month that would exceed the cap is split until it fits. A run that stops can be continued with `--resume` and the same `--out`. ## Options | Option | What it does | | ------------------------------------------ | ----------------------------------------------------------------------------------- | | `--api-key ` | Account API key. `OPSGENIE_API_KEY` in the environment works too. | | `--region us\|eu` | API region. Default `us`. | | `--out ` | Output folder. Default `./opsgenie-export-`. | | `--since ` | Start of alert and incident history. Default: the oldest record. | | `--only ` / `--skip ` | Limit the sections. | | `--with-alert-notes` / `--with-alert-logs` | Fetch notes and the activity log of every alert (one extra request per alert each). | | `--include-secrets` | Keep integration API keys and webhook credentials in the output. | | `--resume` | Skip history months already completed in `/state.json`. | | `--concurrency ` | Parallel per-alert detail requests, 1–10. Default 3. | | `--dry-run` | Validate the key and print what would be exported. | Sections: `account, users, teams, schedules, on-calls, escalations, integrations, policies, maintenance, heartbeats, services, forwarding-rules, roles, alerts, incidents`. ```bash theme={null} # Configuration only — a few hundred requests, a minute or two npx @usewhawit/opsgenie-export --api-key $KEY --skip alerts,incidents # Two years of alerts with their notes, into a folder you can resume npx @usewhawit/opsgenie-export --api-key $KEY --only alerts --since 2025-04-01 \ --with-alert-notes --out ./opsgenie-backup --resume ``` ## What it cannot get * **Audit logs and reports** — not exposed by the API. * **Alert attachments** — listed as download links, not fetched. * **Integration secrets by default** — redacted unless you pass `--include-secrets`. They do not transfer anyway: every integration is re-created and re-keyed on the destination side. * **Anything a team integration key cannot read** — use an account key. ## How long it takes Opsgenie throttles by request domain, and the configuration domain (users, teams, schedules, policies) has the lowest limit. The tool keeps a floor of 120 ms between requests, retries `429` and `5xx` with exponential backoff and honours `Retry-After`. Configuration takes roughly one request per user, team, schedule and integration. History is 100 alerts per request; `--with-alert-notes` and `--with-alert-logs` add one request per alert each, which is what turns a large account into hours rather than minutes. Compare the counts in `manifest.json` with what the Opsgenie UI shows before you rely on the folder, and run the export again close to your cutover date: overrides and notification rules keep changing until the last day. ## Next Recreate schedules, rotations and escalation policies in WHAWIT from the same API key. The cutover playbook: run both systems in parallel, then switch.