Connect Paessler PRTG

Written By Christopher Scaminaci

Last updated 6 days ago

PRTG is Paessler's network and infrastructure monitoring platform. It watches servers, switches, firewalls, printers, cloud services and anything else that answers a check, and it tells you what is up, what is down, and what has been getting slowly worse. Unlike most tools StackJack connects to, PRTG runs on your own equipment: PRTG Network Monitor and PRTG Enterprise Monitor install on a machine you own, and PRTG Hosted Monitor is the version Paessler runs for you. StackJack talks to it through the PRTG API.

Connecting PRTG to StackJack gives your AI assistant a family of prtg_ MCP tools — MCP (Model Context Protocol) tools are the standardized commands an AI assistant can call through StackJack. With them, your AI can:

Answer "what is broken right now" — list the sensors currently in alarm, count how many objects sit in each state across the whole server, and narrow the same question to one site or one customer
Look up any part of the tree — probes, groups, devices, sensors and channels, searched with PRTG's own filter language
Read history — recent measurements for a sensor, the standard graph windows, and arbitrary date ranges for a channel
Report on the estate — the object tree, device templates, sensor kinds, the server version and its license
Handle alarms (on Pro plans) — acknowledge one alarm with a note, or acknowledge everything matching a filter
Run maintenance (on Pro plans) — pause and resume sensors, devices, groups and probes, individually or by filter
Manage the tree (on Pro plans) — create, edit, clone, move and delete objects, and run auto-discovery to find new devices and sensors
Administer accounts (on Pro plans) — users and user groups, plus API key and password tools held behind separate permissions

The three PRTG words that do not mean what you expect

Getting these wrong will send your AI looking in the wrong place, so they are worth thirty seconds:

A probe is a piece of PRTG software — a collector process that performs the monitoring and reports back to your PRTG server. It is not a network device. A remote site usually gets its own probe.
A sensor is one thing being monitored on a device, such as its ping response or one disk volume. A single server commonly has a dozen sensors. PRTG licenses are counted in sensors.
A channel is one measurement inside a sensor, such as the download speed inside a bandwidth sensor. Channels are where the numbers live.

The tree runs probe, then group, then device, then sensor, then channel. Settings flow downward, so a change made on a group reaches everything beneath it.

How StackJack authenticates to PRTG

PRTG uses an API key that you create inside your own PRTG server. There is no client ID, no second secret, and nothing that expires on a schedule.

Two separate things decide what that key can do, and both matter:

The rights of the PRTG user who created it. A key can never reach more than its owner can. A key made by a user who only sees one customer's group stays limited to that group.
The access level chosen for the key, which is Read, Acknowledge, Write or Full. Read allows reporting only. Acknowledge adds alarm acknowledgement. Write adds configuration changes such as pausing, scanning and editing. Full adds account and API key management.

If a tool is refused, check both. Raising the access level cannot widen a key whose owner is restricted.

Steps

Sign in as the right PRTG user. Choose a user who can see every group and device you want StackJack to work with. On a shared PRTG server that usually means an administrator.
Open API Keys — go to Setup, then Account Settings, then My Account, and find the API Keys section.
Add a key and choose its access level. Name it so a later review can tell what holds it, and pick the lowest level that covers what you want StackJack to do.
Copy the key immediately. PRTG shows it once. Treat it as a password — anyone holding it can act on your monitoring as that user.
Find your PRTG server address — the address you reach PRTG at from the internet, including the port of the PRTG Application Server (1616 by default, see below) unless a reverse proxy publishes it on the standard port, such as https://prtg.example.com:1616 or https://prtg.example.com. Enter the address only; StackJack adds the rest of the path itself.
Paste both into StackJack. Open Connectors, choose Paessler PRTG, enter the server address and the API key, and run a Test Connection.

Your PRTG server has to be reachable

Because PRTG runs on your equipment, StackJack has to reach it from the internet. That means it must be published at a public address with a certificate from a public certificate authority. A private or internal-only address will not work, and neither will a self-signed certificate. If your PRTG core still serves the certificate it installed with, replace it before connecting.

A custom port is fine. A reverse proxy that publishes PRTG under a path prefix is also fine — enter the address exactly as you reach it.

StackJack talks to PRTG's API v2, which is served by the PRTG Application Server, a separate service that PRTG installs when the new UI and API v2 are activated (activated by default on new installs from PRTG 25.2; older cores activate it under Setup). It listens on its own port: 1616 for HTTPS by default, 1615 for HTTP. The classic web interface port (443 or 8443 on most installs) does not serve API v2, so a key that PRTG accepts can still be refused through that port. Enter the address and port that reach the Application Server, or publish that service through your reverse proxy.

What to know before your AI uses this connector

Pausing hides outages

This is the most important thing about the PRTG connector, and it is the opposite of how "pause" usually reads. Pausing a sensor, device, group or probe suppresses both the checks and the alerting for everything beneath it. A real failure during a pause produces no alert and no red on the map — a paused estate and a healthy estate look the same at a glance.

The filtered forms make that reach much further, because one call can pause everything matching a filter. Silencing an entire site is a single tool call.

Every pause tool requires the Pro tier and has a matching resume tool. Pausing your monitoring is marked as an ordinary write rather than a destructive one — nothing is destroyed and the resume puts it back — so an AI application that only pauses on destructive actions will not pause on these. Tell your agent to confirm what a filter matches before it pauses in bulk. Pausing a PRTG user account is different: that locks a person out, so it is marked destructive. See Destructive tools and confirmation for what your AI application does with those markings.

Long lists come back one page at a time

A list tool returns one page. PRTG's own maximum page is larger than an AI agent can usefully read, so StackJack limits how many objects one call may ask for. To walk a long list, your agent moves the starting position forward by the size of the page it just read and asks again.

By default a list tool hands back exactly what PRTG sent. If your agent needs to know how many objects exist in total, or whether another page is waiting, it can ask for the totals alongside the page — that is a choice it makes per call, and it changes nothing about what the tools return otherwise.

Scans and auto-discovery send real traffic

A scan asks your PRTG probe to check something immediately rather than waiting for its schedule, so it puts real traffic on your network aimed at real devices. Auto-discovery goes further: it sweeps a range, then creates sensors, which start polling right away and count against your PRTG sensor license. Check your remaining sensor allowance before running a discovery across a group.

Deletes remove history, and PRTG has no undo

Deleting a device, group or probe removes every sensor beneath it and every measurement those sensors ever recorded. There is no recycle bin and no undo. Moving an object is gentler but not harmless: PRTG inherits settings from the parent, so re-parenting an object can quietly change how it is monitored.

Credential and password tools are held separately

Creating an API key, changing one, deleting one, and the two password-reset tools each sit behind their own permission, apart from ordinary user administration. You can allow your AI to manage users while withholding all five, and we suggest you do unless you have a specific reason not to — one of them emails a real person a password-reset link, and editing an existing key can widen what that key is allowed to reach.

Some tools use a part of the API Paessler calls experimental

Creating, updating and deleting objects reaches endpoints Paessler labels experimental and says may change between PRTG releases. They are included because they are the only way to manage the object tree through the API, and each tool says so in its own description. Reading your monitoring — status, alarms, inventory and history — does not depend on them.

Plans and limits

Read tools are available on the Free tier. Everything that acknowledges, pauses, scans, edits, deletes or administers accounts is Pro. Business reaches the same tools as Pro and differs by monthly call quota.

See the generated Paessler PRTG tool reference for the current inventory, plan assignment, input schemas, and destructive-action labels.

Paessler does not publish a request limit for this API, so the pacing StackJack applies exists to protect your server rather than to satisfy the vendor. Your PRTG core is a single machine doing real monitoring work, and a large report should not compete with it. Big reads are spread out instead of arriving all at once. Pacing is not a guarantee: retries are bounded, so a wide enough read can still time out — narrow it rather than repeating it, and check whether a write landed before you retry it (see Retrying a failed or timed-out write).

Several customers

Every customer has their own PRTG server. Add one connection per customer from the connector's card, name it after the customer, and your AI names it on each call. Omit the name and the call runs against your default connection. Pin an endpoint to one connection when an AI should never reach past a single customer. See Several connections of one connector.

Troubleshooting

"PRTG rejected the API key" — two causes. Either the key was deleted or replaced (PRTG keys do not expire, so a key that stops working almost always stopped on purpose: create a new one, paste it into StackJack, and run a Test Connection), or the address reaches PRTG's classic web interface and not the PRTG Application Server. The classic interface refuses an otherwise valid key with the same message, so on a core installed before PRTG 25.2 activate the new UI and API v2 under Setup, then point the connector at the Application Server port (1616 by default; see How StackJack authenticates to PRTG above).

"PRTG accepted the key but refused this operation" — the key's reach, not a fault. Check the access level on the key and the PRTG rights of the user who created it. A Read key is refused for every write tool; a key owned by a restricted user is refused for anything outside that user's groups.

StackJack cannot reach the server at all — work through the address first. It has to be reachable from the internet, over HTTPS, with a certificate from a public authority. A self-signed certificate is the single most common cause, and it fails the same way an unreachable server does.

A device shows no sensors — check whether the device or its parent is paused before assuming discovery failed. An inherited pause from a group or a probe looks exactly like an empty device.

Auto-discovery found nothing — confirm the probe that owns the target range is connected, and that the device credentials stored in PRTG are the ones the target accepts. Discovery runs from the probe, so a probe that is disconnected or on the wrong side of a firewall finds nothing and reports no error you would notice.

A filter returns nothing — PRTG filters are a small query language rather than a list of fields, and an unrecognized term returns an empty result instead of an error. Ask for a plain list first, confirm the objects are there, then add the filter.

Paessler PRTG tools

prtg_ · 100 tools · Free 43 · Pro 57

Sensors & Alarms

Tool	What it does
`prtg_acknowledge_sensor_alarm` Pro · Write	Acknowledge one sensor's alarm, which stops PRTG re-notifying about it while leaving the sensor MONITORING NORMALLY.
`prtg_acknowledge_sensor_alarms_filtered` Pro · Write	Acknowledge the alarms of EVERY sensor matching a filter — the bulk form of prtg_acknowledge_sensor_alarm.
`prtg_create_sensor` Pro · Write	Create a new sensor on a device.
`prtg_delete_sensor` Pro · Destructive	Delete a sensor AND EVERY CHANNEL UNDER IT, permanently.
`prtg_get_object_status_summary` Free · Read-only	Count the sensors in each state UNDER one probe, group or device — the scoped form of prtg_get_sensor_status_summary.
`prtg_get_sensor` Free · Read-only	Get one sensor's current metrics and settings by id.
`prtg_get_sensor_status_summary` Free · Read-only	Count the sensors in each state across the WHOLE PRTG server — up, down, warning, paused, unusual, unknown.
`prtg_list_alarms` Free · Read-only	List the sensors currently in an alarm state — the fastest answer to "what is broken right now".
`prtg_list_sensor_measurements` Free · Read-only	List one sensor's recent channel measurements — the recorded values for each of its channels.
`prtg_list_sensors` Free · Read-only	List sensors on the PRTG server, with the filter language for narrowing.
`prtg_metascan_device` Pro · Destructive	Ask a device which instances of one sensor kind exist on it — which network interfaces, which disk volumes, which services — so the results can be turned into sensors with prtg_create_sensor.
`prtg_pause_sensor` Pro · Write	Pause one sensor.
`prtg_pause_sensors_filtered` Pro · Write	Pause EVERY sensor matching a filter.
`prtg_resume_sensor` Pro · Write	Resume one paused sensor, so PRTG starts checking and alerting on it again.
`prtg_resume_sensors_filtered` Pro · Write	Resume EVERY paused sensor matching a filter — the bulk form of prtg_resume_sensor, and the normal way to end a maintenance window.
`prtg_scan_sensor_now` Pro · Destructive	Trigger an immediate check of one sensor instead of waiting for its next scheduled interval.
`prtg_scan_sensors_now_filtered` Pro · Destructive	Trigger an immediate check of EVERY sensor matching a filter.
`prtg_update_sensor` Pro · Write	Update a sensor's settings.

Devices

Tool	What it does
`prtg_create_device_in_group` Pro · Write	Create a new device inside a group.
`prtg_create_device_on_probe` Pro · Write	Create a new device directly on a probe rather than inside a group.
`prtg_delete_device` Pro · Destructive	Delete a device AND EVERY SENSOR AND CHANNEL UNDER IT, permanently.
`prtg_get_device` Free · Read-only	Get one device's metrics and settings by id — host address, status, scanning interval, parent group and sensor roll-up.
`prtg_list_device_icons` Free · Read-only	List the device icon names the PRTG interface offers.
`prtg_list_device_templates` Free · Read-only	List the device templates auto-discovery can apply.
`prtg_list_devices` Free · Read-only	List the devices PRTG monitors — servers, switches, firewalls, URLs.
`prtg_pause_device` Pro · Write	Pause one device.
`prtg_pause_devices_filtered` Pro · Write	Pause EVERY device matching a filter, and with it every sensor on those devices.
`prtg_resume_device` Pro · Write	Resume one paused device, restoring monitoring and alerting for every sensor on it.
`prtg_resume_devices_filtered` Pro · Write	Resume EVERY paused device matching a filter — the bulk form of prtg_resume_device, and the normal way to end a maintenance window across a site.
`prtg_run_device_autodiscovery` Pro · Destructive	Run auto-discovery against one device: PRTG probes it, works out what can be monitored, and CREATES SENSORS for what it finds.
`prtg_scan_device_now` Pro · Destructive	Trigger an immediate check of EVERY sensor on one device.
`prtg_scan_devices_now_filtered` Pro · Destructive	Trigger an immediate check of every sensor on EVERY device matching a filter.
`prtg_update_device` Pro · Write	Update a device's settings.

Groups

Tool	What it does
`prtg_create_group_in_group` Pro · Write	Create a new group inside an existing group — how a site or department is added under a customer's top-level group.
`prtg_create_group_on_probe` Pro · Write	Create a new TOP-LEVEL group directly on a probe — the usual first step when onboarding a new customer into a shared PRTG server.
`prtg_delete_group` Pro · Destructive	Delete a group AND EVERY GROUP, DEVICE, SENSOR AND CHANNEL UNDER IT, permanently.
`prtg_get_group` Free · Read-only	Get one group's metrics and settings by id — status, scanning interval, parent, and the settings its devices inherit.
`prtg_list_groups` Free · Read-only	List the groups on the PRTG server.
`prtg_pause_group` Pro · Write	Pause one group.
`prtg_pause_groups_filtered` Pro · Write	Pause EVERY group matching a filter, and with it every device and sensor beneath those groups.
`prtg_resume_group` Pro · Write	Resume one paused group, restoring monitoring and alerting for every device and sensor beneath it.
`prtg_resume_groups_filtered` Pro · Write	Resume EVERY paused group matching a filter — the bulk form of prtg_resume_group, and the normal way to end a site-wide maintenance window.
`prtg_run_group_autodiscovery` Pro · Destructive	Run auto-discovery across EVERY device in a group: PRTG probes each one, works out what can be monitored, and CREATES SENSORS for what it finds.
`prtg_scan_group_now` Pro · Destructive	Trigger an immediate check of EVERY sensor in one group.
`prtg_scan_groups_now_filtered` Pro · Destructive	Trigger an immediate check of every sensor in EVERY group matching a filter.
`prtg_update_group` Pro · Write	Update a group's settings.

Probes

Tool	What it does
`prtg_delete_probe` Pro · Destructive	Delete a probe AND EVERY GROUP, DEVICE, SENSOR AND CHANNEL UNDER IT, permanently.
`prtg_get_probe` Free · Read-only	Get one probe's metrics and settings by id — connection state, version, and the settings every group and device beneath it inherits.
`prtg_get_probe_network_info` Free · Read-only	Get the network interfaces and IP addresses of the MACHINE a probe runs on.
`prtg_list_probes` Free · Read-only	List the probes attached to the PRTG core server.
`prtg_pause_probe` Pro · Write	Pause one probe.
`prtg_pause_probes_filtered` Pro · Write	Pause EVERY probe matching a filter.
`prtg_resume_probe` Pro · Write	Resume one paused probe, restoring monitoring and alerting for every group, device and sensor beneath it.
`prtg_resume_probes_filtered` Pro · Write	Resume EVERY paused probe matching a filter — the bulk form of prtg_resume_probe.
`prtg_scan_probe_now` Pro · Destructive	Trigger an immediate check of EVERY sensor a probe collects.
`prtg_scan_probes_now_filtered` Pro · Destructive	Trigger an immediate check of every sensor on EVERY probe matching a filter.
`prtg_update_probe` Pro · Write	Update a probe's settings.

Objects & Schemas

Tool	What it does
`prtg_clone_object` Pro · Write	Clone an object to a new parent — the fast way to stand up a second site from a working one, or to copy a device's whole sensor set onto a new host.
`prtg_get_object_counts` Free · Read-only	Count how many objects of each kind exist on the server — probes, groups, devices, sensors, channels.
`prtg_get_object_create_schema` Free · Read-only	Get the OpenAPI schema for the CREATE body of one object kind on one parent — exactly which properties are required and allowed.
`prtg_get_object_read_schema` Free · Read-only	Get the OpenAPI schema describing what a READ of one specific object returns — which properties exist and what type each one is.
`prtg_get_object_update_schema` Free · Read-only	Get the OpenAPI schema for the UPDATE body of one specific object — exactly which properties prtg_update_sensor, prtg_update_device, prtg_update_group or prtg_update_probe will accept for that id.
`prtg_get_setting_lookup` Free · Read-only	List the legal values of one named setting — the only way to find out what an enumerated PRTG setting will accept before writing it.
`prtg_list_creatable_object_kinds` Free · Read-only	List the object kinds that can be created on a given parent — which sensor kinds a device accepts, or whether a probe accepts groups and devices.
`prtg_move_object` Pro · Destructive	Move a live monitoring object to a new parent, or reorder it among its siblings.
`prtg_search_objects` Free · Read-only	Search across EVERY object kind at once — probes, groups, devices, sensors and channels — using the PRTG filter language.

Auto-Discovery

Tool	What it does
`prtg_create_autodiscovery_task_on_probe` Pro · Write	Create a saved auto-discovery task on a probe.
`prtg_delete_autodiscovery_task` Pro · Destructive	Delete a saved auto-discovery task permanently.
`prtg_get_autodiscovery_task` Free · Read-only	Get one saved auto-discovery task by id — its target range, the device templates it applies, its schedule and the probe that would run it.
`prtg_list_autodiscovery_tasks` Free · Read-only	List the SAVED auto-discovery tasks — the stored definitions of what to scan and which device templates to apply.
`prtg_start_autodiscovery_task` Pro · Destructive	Start a saved auto-discovery task now.
`prtg_update_autodiscovery_task` Pro · Write	Update a saved auto-discovery task.

Channels & Time Series

Tool	What it does
`prtg_get_channel` Free · Read-only	Get one channel's last value and settings by id — the unit, the display formatting, and the warning and error limits that decide when its parent sensor changes state.
`prtg_get_sensor_graph_set` Free · Read-only	Get historic channel data for one sensor over one of PRTG's four PRESET windows — the same data its interface graphs show.
`prtg_get_sensor_timeseries` Free · Read-only	Get historic channel data for one sensor over an ARBITRARY time range — the tool for 'what did this look like during Tuesday's outage' or 'show me last month for the capacity review'.
`prtg_list_channel_measurements` Free · Read-only	List the last measurement of many channels at once, each with a reference back to its channel — a roll-up for building a dashboard view without a call per sensor.
`prtg_list_channels` Free · Read-only	List channels.

Users & User Groups

Tool	What it does
`prtg_get_user` Free · Read-only	Get one PRTG user account by id — its login name, email address, user group memberships and account state.
`prtg_get_user_group` Free · Read-only	Get one PRTG user group by id — its members and the access rights they inherit from it.
`prtg_list_user_groups` Free · Read-only	List the PRTG user groups.
`prtg_list_users` Free · Read-only	List the user accounts on the PRTG server — who can sign in, which user groups they belong to and whether their account is active or paused.
`prtg_pause_user` Pro · Destructive	Pause a PRTG user account.
`prtg_pause_users_filtered` Pro · Destructive	Pause EVERY user account matching a filter — a bulk ACCOUNT LOCK.
`prtg_resume_user` Pro · Write	Resume a paused PRTG user account, letting that person sign in again and re-enabling any API keys they own.
`prtg_resume_users_filtered` Pro · Write	Resume EVERY paused user account matching a filter — the bulk form of prtg_resume_user.

API Keys & Password Reset

Tool	What it does
`prtg_create_api_key` Pro · Destructive	Mint a new API key for a PRTG user account.
`prtg_delete_api_key` Pro · Destructive	Delete an API key permanently, revoking it immediately.
`prtg_find_api_key` Free · Read-only	Look up one API key's record by its VALUE or by its id — which user owns it, what it is named and what access level it carries.
`prtg_list_user_api_keys` Free · Read-only	List the API keys belonging to one user account — their names, descriptions and access levels.
`prtg_reset_password` Pro · Destructive	Set a new password for a PRTG user account using the token from the reset email.
`prtg_trigger_password_reset` Pro · Destructive	Start the password reset process for a PRTG user account.
`prtg_update_api_key` Pro · Destructive	Edit an existing API key's name, description or access level.

System & Lookups

Tool	What it does
`prtg_check_health` Free · Read-only	Check whether the PRTG application server is up.
`prtg_get_license_info` Free · Read-only	Get the PRTG license — the edition, the sensor allowance and how much of it is used.
`prtg_get_lookup_definition` Free · Read-only	Get one lookup definition by id — the full mapping from raw sensor values to status text and sensor state.
`prtg_get_public_settings` Free · Read-only	Get the settings PRTG exposes to unauthenticated callers — the small set its own sign-in page needs, such as the server name and which login methods are offered.
`prtg_get_version` Free · Read-only	Get the PRTG core server and application server versions.
`prtg_list_libraries` Free · Read-only	List the PRTG libraries — saved custom views that gather objects from anywhere in the tree into one arrangement, often built per customer for a status wall.
`prtg_list_lookup_definitions` Free · Read-only	List the lookup definitions on the server.
`prtg_list_running_autodiscoveries` Free · Read-only	List the auto-discoveries running on the server RIGHT NOW, with their progress.

Still need help? Ask the team