Configuration Export and Import
Organization admins and owners can export the organization's alert rules, automations and saved Data Explorer views as one JSON file, and import such a file into the same organization or another one. Use it to keep a backup of a setup, to restore it, or to copy it to another organization. To move a setup from another RMM tool, the card can also draft a configuration file with AI from that tool's screenshots, PDFs, export files or text (see Importing from another tool, and Moving a setup from another tool, step by step for the whole process).
Where to find it
Open Configuration, then the Workspace tab. The Configuration Export and Import card holds the export button, the file picker and the import report. Export and import act on the organization you have selected. A member who is not an admin or owner of that organization gets the message “Only organization admins and owners can export or import configuration.”
What the file contains
| Section | Included | Left out |
|---|---|---|
| Alert rules | Every alert rule of the organization: name, description, severity, evaluation interval, conditions, tags, and the name and type of each linked notification channel. | Notification channel settings, because they can hold credentials. |
| Automations | Every automation rule of the organization: name, description, trigger, actions and their settings, target scope and tag, schedule, minimum severity, trigger conditions and retry settings. | Webhook URLs, webhook headers and webhook body templates, because they can hold credentials. The file lists, for each rule, which of them were left out. |
| Reports | Your own saved Data Explorer views and the views shared with the organization: name, description, chart type and query. | Other members' private views. |
The file starts with "format": "etducky.config" and "version": 1. An import refuses a file without that format, or with a newer version than the server reads.
Smart Reports questions are not saved as definitions, so they are not part of the file. Saved Data Explorer views are.
Importing a file
- Choose the file with the file picker.
- Click Preview import. The server checks every item against the organization and shows what the import would add and a report of anything that would be skipped or changed. Nothing is written.
- Click Import to add the items.
- Review each imported rule, fill in what the report names, and enable it.
An import never turns anything on. Imported alert rules and automations are disabled. Imported saved views are private to the member who imported them, and can be shared from Data Explorer.
A disabled automation does not run. Schedules and event triggers skip it, and Run Now answers “This automation is disabled. Run Now runs only enabled automations. Enable it in the rule editor, then run it again.”
What the import changes
| Situation | Result |
|---|---|
| An alert rule names a notification channel | The rule is linked to the channel of the same name and type in this organization. When none exists, the rule is imported without that link and the report says so. |
| An alert rule fails the rule editor's checks (severity, an evaluation interval outside 10 to 3600 seconds, missing or invalid conditions) | The rule is not imported. The report gives the reason. |
| A name already exists in this organization | The item is imported anyway, and both are kept. The report says so for alert rules and automations. |
| An automation step runs a script that is not in this organization and is not a built-in script | The script is cleared from the step. Choose a script before enabling the rule. |
| An automation step sends to an alert channel that is not in this organization | The channel is cleared from the step. |
| An automation targets a tag that does not exist in this organization | The tag name is kept, and the report says the tag does not exist yet. |
| An automation assigns tickets to a member | The assignee is kept. A rule whose assignee is not a member of the organization fails when it runs. |
| An automation step in the file sets a webhook URL (an ET Ducky export never does, but a converted or edited file can) | The URL is kept, and the report names the host it points at. Check it before enabling the rule. |
| An automation's webhook URL, headers or body template were left out of the export | The report reminds you to enter them before enabling the rule. |
| An automation has no name, trigger or action, or a target scope other than triggering agent, all agents or tagged | The automation is not imported. |
| A saved view has no query or its query names no data source | The view is not imported. |
| Saved views on an organization without a paid subscription | Saved views are not imported. Data Explorer saved views need a paid subscription. |
| The organization already has 50 saved views | The remaining views are not imported. |
Importing from another tool
The Import from another tool section of the card drafts a configuration file from another RMM tool's alert rules, monitors, automations and reports. The draft is not imported until you preview it and click Import, and it imports the same way as any other file: rules disabled, views private. Drafting is for organization admins and owners, and it is part of every paid plan, including plans with agent seats only.
- Optionally choose the other tool in the list. Choosing Datto RMM also gives the AI Datto's monitor types, alert priorities and policy terms, so monitors map more reliably.
- Add up to 5 files, paste text, or both. Files can be screenshots (PNG, JPEG or WebP), PDFs, or export files (XML, JSON, CSV or plain text, in UTF-8, in UTF-16 with a byte order mark, or in Latin-1, also called ISO-8859-1). ZIP archives are refused: extract the files inside and upload those.
- Click Draft with AI. The draft is made on the server while the card shows its progress: waiting for a drafting slot, drafting, or waiting to ask the AI provider again. Keep the page open. When the draft is done, the card shows what the AI could not map, the fields it left blank and its notes, then previews the draft as described in Importing a file.
- Click Import, or Download draft to keep the draft file and edit it first.
| In the files | In the draft |
|---|---|
| Webhook URLs, webhook headers and webhook body templates | Always blank. Each one is listed under “Left blank: fill in after import”. |
| Any other value that looks like a key, token or password, a URL that carries a password or key, or an incoming-webhook URL | Blank, and listed under “Left blank: fill in after import”. |
| Scripts, alert channels, ticket integrations and ticket assignees | Never set. Choose them in the rule editor after import. A note names each automation whose run_script step needs a script. |
| Notification channels of alert rules | Not linked. Link channels in the alert rule after import. |
| A condition on a metric ET Ducky does not read, an automation trigger that ET Ducky does not fire (Alert Fired, Agent Went Offline, Agent Came Online), or an action ET Ducky does not have | The item is listed under “Not mapped” and left out of the draft. |
| A schedule | A standard 5-field cron expression that runs in UTC (see Automations), converted from the other tool's time zone when the files show it. A schedule cron can't express, such as seconds, the last day of the month or the second Tuesday, is listed under “Not mapped”. |
| A report | A saved Data Explorer view over the last 7 days, when its query passes Data Explorer's checks. Views on a data source that needs a privilege are listed under “Not mapped”. |
| Anything that fails the checks an import applies, or that the AI reports it could not map | Listed under “Not mapped” with the reason, and left out of the draft. |
| Things the AI could not see, such as collapsed panels, script bodies, or schedules on another tab | Listed under “Notes from the AI”. |
The AI is told to treat everything in the files as data and to ignore instructions written in them. Its answer is then checked item by item, so only items that pass the checks reach the draft.
Moving a setup from another tool, step by step
Use these steps the first time you move alert rules, monitors, automations or reports into ET Ducky.
1. Prepare ET Ducky
- Install agents on at least a few of the devices the rules are for, so you can test them after import (see Agents).
- Create the notification channels, tags and scripts the rules will need. An import links an alert rule to a channel of the same name and type, keeps tag names, and keeps scripts only when they already exist in the organization or are built in.
2. Collect material from the other tool
- An export file is the most complete source when the other tool offers one. Use XML, JSON, CSV or plain text.
- Otherwise take screenshots of each rule, monitor, automation or report with its settings expanded, so the conditions, thresholds, schedule and actions are visible. Settings in collapsed panels, on other tabs or inside script bodies are not seen.
- Include the time zone a schedule runs in when the other tool shows it. Without it, check each converted schedule after import.
- Remove passwords, API keys and tokens from export files before you upload them where you can. The draft leaves values that look like secrets blank in any case.
- Group the material by kind or by policy. A draft takes at most 5 files and holds at most 100 items of each kind, and an organization can make as many drafts as it needs.
3. Draft, preview and import
- Open Configuration, then Workspace, and find Import from another tool in the Configuration Export and Import card.
- Choose the other tool if it is listed, add the files or text, and click Draft with AI.
- Read the “Not mapped”, “Left blank: fill in after import” and “Notes from the AI” lists. Keep a copy: click Download draft if you want the file.
- Check the preview, then click Import. Repeat for the next group of material.
4. After the import
Everything imported is off until you turn it on. For each imported alert rule and automation:
- Open it in the rule editor and compare it with the original: conditions, thresholds, severity, evaluation interval and target scope.
- Link the notification channels of alert rules.
- Choose the script for each run_script step, and the alert channel, ticket integration or assignee for steps that need one.
- Enter webhook URLs, headers and body templates listed as left blank.
- Check each schedule. Schedules run in UTC.
- Create any tag the import report says does not exist yet, and add it to the devices the rule is for.
- Enable the rule. To try an automation on a few devices first, set its target scope to a tag that only test devices carry, then widen it (see Automations).
Rebuild by hand the items listed under “Not mapped”, or leave them out where ET Ducky covers the need another way. Imported saved views are private to you; share them from Data Explorer when they are ready.
Limits and records
- A file may carry at most 500 alert rules, 500 automations and 500 saved views, and may be at most 4 MB.
- Every export and every import (not a preview) is recorded in the organization's audit log with the number of items.
- An AI draft takes at most 5 files of up to 4 MB each, 10 MB in all, and 120,000 characters of text. The file type is read from the file's content, not its name, and other types are refused.
- Files and text for a draft are sent to the organization's AI provider for that one draft: the organization's own key when BYOK is set, otherwise ET Ducky's provider. ET Ducky does not store them. The ET Ducky server that receives them keeps them in its memory only until the draft is finished. The finished draft is kept for an hour so the card can show it, then deleted.
- An organization without a paid plan gets the answer “AI drafts are part of every paid ET Ducky plan, including plans with agent seats only.” There is no limit on how many drafts an organization makes.
- Each draft uses queries from the organization's AI query pool, metered by tokens, unless the organization's own BYOK key makes it (see AI Usage). When the pool is used up and the organization has no key of its own that works, the draft fails with a sentence that says how to get more queries or add a key.
- When the organization's own key refuses or fails a draft, ET Ducky's provider makes the draft instead. The card then says that the draft was charged to the organization's AI queries.
- When the organization's own key does not accept PDF files (for example an Azure OpenAI deployment without PDF input), ET Ducky reads the text of each PDF on its own server and sends that text with the same key instead. The key is not marked as failing, nothing is charged to the AI queries, and the card says the PDF's text was sent; images, tables and layout in the PDF are not sent. A PDF with no text ET Ducky can read, such as a scanned page, ends the draft with a sentence asking for screenshots or pasted text.
- When the AI provider does not answer in time or limits requests, the draft is tried again up to three more times, after 20, 60 and 120 seconds, and the card shows the wait. When the organization's own key does not answer in time and the organization has no AI queries left for ET Ducky's provider to take over, the draft fails at once instead of waiting. When the card gets no answer from ET Ducky while it checks on a draft, for example during a network interruption, it says it is waiting and checks again; the draft keeps running on the server.
- Each ET Ducky server makes a few drafts at a time, at most two of one organization's at once, and other drafts wait their turn; there is no limit on how many wait. When a server is holding as many waiting uploads as it can, the card sends the draft again after a short wait. If the server holding a draft's files stops before the draft is done, the draft fails, because the files were never stored.
- When a draft fails, the card shows a Retry button. It starts a new draft from the same files, vendor and text, which the browser still holds; ET Ducky's servers never stored them. If you changed the chosen files or the text since, or left the page, the card asks you to choose the files again and click Draft with AI.
- A draft holds at most 100 items of each kind. Upload fewer items at a time for more.
- Every AI draft is recorded in the audit log with the number of files and items, never file names or contents.