Notification Settings

Nezha Monitor can monitor server load, CPU, memory, disk, traffic, monthly traffic, process count, and connection count, and send notifications when user-defined thresholds are reached.

Flexible Notification Methods

In panel messages, #DATETIME# represents the timestamp when the event occurred and is automatically replaced with the actual time when a notification is triggered.
#NEZHA# is the panel message placeholder and is automatically replaced with the actual message content when a notification is triggered.
Request body format:
- When the request type is FORM, use key:value format. value can contain placeholders that are automatically replaced during notification.
- When the request type is JSON, simple string replacement is performed and the result is submitted directly to the URL.
The URL can also contain placeholders, which are replaced during the request.
Verify TLS: When enabled, HTTPS notification requests verify the certificate chain and domain. If the notification service uses a self-signed certificate or internal test certificate, you may need to disable this option.
Skip Check: When adding or modifying a notification method, the Dashboard sends a test message by default. After checking “Skip Check”, the configuration is saved directly. This is useful when the notification service is temporarily unreachable but you are sure the parameters are correct.
Format Metric Units: If “Format Metric Units” is enabled, #SERVER.CPU#, #SERVER.MEM#, #SERVER.SWAP#, #SERVER.DISK#, #SERVER.SPEEDIN#, #SERVER.SPEEDOUT#, #SERVER.TRANSFERIN#, and #SERVER.TRANSFEROUT# output human-readable values with units, such as 12.34 %, 1.23 GB, and 10.50 MB/s. If disabled, these placeholders keep their raw numeric values, which is suitable for webhooks that need pure numbers.
Webhook Target Restrictions: Notification URLs only support http and https. For security reasons, the Dashboard rejects notification requests pointing to localhost, private networks, link-local addresses, reserved addresses, multicast addresses, and other non-public targets, and it does not automatically follow redirects. If you need to push to an internal service, receive the notification through a publicly accessible relay first.
Editing Sensitive Fields: The API and admin frontend do not display the URL, request headers, or request body when reading a notification method. Re-enter the URL when editing. Leaving headers or body empty preserves their stored values. To clear stored headers or body, delete and recreate the notification method.

Refer to the following notification examples. You can also customize push methods as needed.

Supported Placeholders

Placeholder	Meaning
`#NEZHA#`	Notification content
`#SERVER.ID#`	Server ID
`#SERVER.NAME#`	Server name
`#SERVER.IP#`	Server IP
`#SERVER.IPV4#`	Server IPv4 address
`#SERVER.IPV6#`	Server IPv6 address
`#SERVER.CPU#`	CPU usage, outputs a percentage when “Format Metric Units” is enabled
`#SERVER.MEM#`	Memory usage, outputs a percentage when “Format Metric Units” is enabled
`#SERVER.SWAP#`	Swap usage, outputs a percentage when “Format Metric Units” is enabled
`#SERVER.DISK#`	Disk usage, outputs a percentage when “Format Metric Units” is enabled
`#SERVER.SPEEDIN#`	Real-time inbound speed, outputs `KB/s`, `MB/s`, etc. when “Format Metric Units” is enabled
`#SERVER.SPEEDOUT#`	Real-time outbound speed, outputs `KB/s`, `MB/s`, etc. when “Format Metric Units” is enabled
`#SERVER.TRANSFERIN#`	Total inbound traffic, outputs `KB`, `MB`, `GB`, etc. when “Format Metric Units” is enabled
`#SERVER.TRANSFEROUT#`	Total outbound traffic, outputs `KB`, `MB`, `GB`, etc. when “Format Metric Units” is enabled
`#SERVER.CPUUSED#`	Raw CPU value
`#SERVER.MEMUSED#`	Used memory in bytes
`#SERVER.MEMTOTAL#`	Total memory in bytes
`#SERVER.SWAPUSED#`	Used swap in bytes
`#SERVER.SWAPTOTAL#`	Total swap in bytes
`#SERVER.DISKUSED#`	Used disk in bytes
`#SERVER.DISKTOTAL#`	Total disk in bytes
`#SERVER.NETINSPEED#`	Raw inbound speed in bytes/second
`#SERVER.NETOUTSPEED#`	Raw outbound speed in bytes/second
`#SERVER.NETINTRANSFER#`	Raw inbound traffic in bytes
`#SERVER.NETOUTTRANSFER#`	Raw outbound traffic in bytes
`#SERVER.LOAD1#`	1-minute load
`#SERVER.LOAD5#`	5-minute load
`#SERVER.LOAD15#`	15-minute load
`#SERVER.TCPCONNCOUNT#`	TCP connection count
`#SERVER.UDPCONNCOUNT#`	UDP connection count

Bark Example

Expand/collapse

Name: Bark
URL composition:
- GET request: /:key/:body, /:key/:title/:body, or /:key/:category/:title/:body
- POST request: /push
Request method:
- GET or POST
Request type:
- GET request: default
- POST request: form

Body:

GET request: empty

POST request:

json

{
  "title": "#SERVER.NAME#",
  "device_key": "xxxxxxxxx",
  "body": "#NEZHA#",
  "icon": "https://xxxxxxxx/nz.png"
}

Slack Example (Contributed by @cantoblanco)

Expand/collapse

Get URL Parameters

Create a Slack App: Create a new app at Slack API.
Add Incoming Webhook: Add Incoming Webhooks on the app feature page and activate them.
Generate Webhook URL: Click “Add New Webhook to Workspace”, select a channel, and authorize it to get the Webhook URL.

Notification configuration:

Name: Slack
URL: https://hooks.slack.com/services/XXXXXXXXX/XXXXXXXXX/XXXXXXXXXXXXXXXXXXXXXXXX
Request method: POST
Request type: JSON
Body:
json
```
{
  "text": "#NEZHA#"
}
```

ServerChan Example

Expand/collapse

Simple example:
- Name: ServerChan
- URL: https://sc.ftqq.com/SCUrandomkeys.send?title=Nezha notification&desp=#NEZHA#
- Request method: GET
- Request type: default
- Body: empty

Advanced example:

Name: ServerChan
URL: https://sc.ftqq.com/SCUrandomkeys.send
Request method: POST
Request type: FORM

Body:

json

{
  "title": "#SERVER.NAME#",
  "desp": "**#NEZHA#\n\nAverage load: \"#SERVER.LOAD1#\",\"#SERVER.LOAD5#\",\"#SERVER.LOAD15#\"\n\n## [Open Dashboard](https://your-dashboard-domain)\n\n![logo](https://raw.githubusercontent.com/naiba/nezha/master/resource/static/brand.svg)"
}

Notification preview:

ServerChan notification preview

Telegram Example (Contributed by @cantoblanco)

Expand/collapse

Get URL Parameters

Get bot token: Chat with @BotFather, send /newbot to create a new bot, and get the token.
Get user ID: Chat with @userinfobot to get your user ID.
Chat with the bot: Send a message to the newly created bot first to ensure it can send messages to you.

Notification configuration:

Name: Telegram
URL: https://api.telegram.org/bot<your-bot-token>/sendMessage?chat_id=<your-user-id>&text=#NEZHA#
Request method: GET
Request type: default
Body: empty

Note: Replace <your-bot-token> and <your-user-id> with actual values.

wxpusher Example

Expand/collapse

Follow your application in advance.

Name: wxpusher
URL: http://wxpusher.zjiecode.com/api/send/message
Request method: POST
Request type: JSON

Body:

json

{
  "appToken": "your appToken",
  "topicIds": [],
  "content": "#NEZHA#",
  "contentType": "1",
  "uids": ["your uid"]
}

Email Notification Example (Using SendCloud, Contributed by @cantoblanco, @erbanku)

Expand/collapse

Note: SendCloud has a daily free email quota. This is only an example.

Get URL Parameters

Register an account:

China SendCloud: Register and log in at SendCloud.
International SendCloud: Register and log in at SendCloud.

Get APIUSER and APIKEY: Get them from account settings.
Set sender and recipient email: Customize sender and recipient email addresses.

Notification configuration:

Name: Email notification

URL (China version):

https://api.sendcloud.net/apiv2/mail/send?apiUser=<APIUSER>&apiKey=<APIKEY>&from=<sender-email>&fromName=Nezha&to=<recipient-email>&subject=Nezha-Notification&html=#NEZHA#

URL (International version):

https://api2.sendcloud.net/api/mail/send?apiUser=<APIUSER>&apiKey=<APIKEY>&from=<sender-email>&fromName=Nezha&to=<recipient-email>&subject=Nezha-Notification&html=#NEZHA#

Request method: POST
Request type: JSON
Header: empty
Body: empty

Note: Replace <APIUSER>, <APIKEY>, <sender-email>, and <recipient-email> with actual values.

DingTalk Group Bot Example

Expand/collapse

Get URL Parameters

Create a bot: Add a bot in DingTalk group settings and select custom keyword mode.
Get Webhook URL: Obtain it after creation.

Notification configuration:

Name: Nezha Monitor Bot
URL: https://oapi.dingtalk.com/robot/send?access_token=xxxxxxxxxxxxxxxxx
Request method: POST
Request type: JSON

Header:

json

{
  "Content-Type": "application/json"
}

Body:

json

{
  "msgtype": "text",
  "text": {
    "content": "Nezha Monitor:\n#NEZHA#"
  }
}

Note: Replace access_token with your bot token, and the content must include your custom keyword.

WeCom Group Bot Example (Contributed by @ChowRex)

Expand/collapse

Configuration Example

Name: WeCom group bot
URL: https://qyapi.weixin.qq.com/cgi-bin/webhook/send?key=YOUR_BOT_KEY
Request method: POST
Request type: JSON

Body:

json

{
  "msgtype": "markdown",
  "markdown": {
    "content": "# Nezha Notification\n\n#NEZHA#\n\n> **Name**: #SERVER.NAME#\n> **IP**: #SERVER.IP#\n> **CPU**: #SERVER.CPU#%\n> **Memory**: #SERVER.MEM#%\n> **Disk**: #SERVER.DISK#%\n"
  }
}

Notification preview:

WeCom notification preview

Feishu Group Bot Example (Contributed by @eya46)

Expand/collapse

Get URL Parameters

Create a bot: Add a custom bot (Webhook) in Feishu group settings.
Get Webhook URL: Obtain it after creation.

Notification configuration:

Name: Nezha Dashboard Bot
URL: https://open.feishu.cn/open-apis/bot/v2/hook/xxxxxxxxxxxxxxxxx
Request method: POST
Request type: JSON

Body:

json

{
  "msg_type": "text",
  "content": {
    "text": "#NEZHA#\n#DATETIME#"
  }
}

Matrix Notification Example

Expand/collapse

Parameter Description

Variable replacement: Replace these variables with actual values:
- $YOUR_HOME_SERVER: Matrix server address
- $YOUR_NEZHA_URL: Your Nezha Dashboard URL
- $YOUR_MATRIX_USERNAME: Matrix username
- $YOUR_MATRIX_PASSWD: Matrix password
- $YOUR_MATRIX_TOKEN: Matrix access token
- $ROOM_ID: Matrix room ID

Retrieve Access Token

Use this command to get YOUR_MATRIX_TOKEN:

bash

curl -XPOST -d '{"type": "m.login.password", "identifier": {"user": "$YOUR_MATRIX_USERNAME", "type": "m.id.user"}, "password": "$YOUR_MATRIX_PASSWD"}' "https://$YOUR_HOME_SERVER/_matrix/client/r0/login"

Notification Configuration

Name: Matrix
URL: https://$YOUR_HOME_SERVER/_matrix/client/r0/rooms/$ROOM_ID/send/m.room.message
Request method: POST
Request type: JSON

Header:

json

{
  "Authorization": "Bearer $YOUR_MATRIX_TOKEN"
}

Body:

json

{
  "msgtype": "m.text",
  "format": "org.matrix.custom.html",
  "formatted_body": "<h1><a href=\"$YOUR_NEZHA_URL\" target=\"_blank\">Nezha Dashboard</a></h1><ul><li>Time: #DATETIME#</li><li>Message: #NEZHA#</li></ul>",
  "body": "#NEZHA#"
}

Steps to use:

Replace variables: Replace all variables beginning with $ with your actual values.
Get token: Use the command above to get YOUR_MATRIX_TOKEN.
Configure request: Fill in URL, Header, and Body according to the configuration.
Send notifications: After configuration, notifications are sent to the specified Matrix room.

Alert Rule Description

Before configuring alert rules, configure notification methods first, then configure notification groups for notification methods in group settings. Alert rules send notifications by notification group.

Basic Rules

type: Select one or more types. A notification is triggered only when all selected types are satisfied at the same time.
- Resource: cpu, gpu, memory, swap, disk
- Network: net_in_speed (inbound speed), net_out_speed (outbound speed), net_all_speed (bidirectional speed), transfer_in (inbound traffic), transfer_out (outbound traffic), transfer_all (bidirectional traffic)
- System: offline (offline monitoring), load1, load5, load15 (load), process_count (process count)
- Connections: tcp_conn_count, udp_conn_count
- Temperature: temperature_max (maximum temperature value)
duration: An integer of at least 3, specifying an observation window of duration × 3 seconds. Evaluation uses time covered by reports. A normal threshold rule triggers when values are outside the allowed range for more than 70% of the window, with sufficient data coverage. Offline rules use a continuous offline window; actual trigger timing varies with the check cadence.
min / max:
- Traffic and speed units are bytes (1KB=1024B, 1MB=1024*1024B).
- Memory, disk, and CPU use percentages (0-100).
- Offline monitoring does not need this item.
cover:
- 0: Monitor all servers. Use ignore to ignore specific servers.
- 1: Ignore all servers. Use ignore to select specific servers for monitoring.
ignore: Used with cover to specify server monitoring strategy, in the format {server ID: true/false}.

Metric rules use distinct fresh reports; the same report is not counted repeatedly. Reports older than 6 seconds, missing GPU data, or zero denominators such as unavailable total memory make the relevant condition unknown. Unknown data neither starts an incident nor sends a recovery; a new window must accumulate once data becomes available again. A compound rule cannot decide an incident or recovery while any applicable condition is unknown. Configure offline alerts separately from CPU, GPU, and other resource alerts.

An offline-only rule still advances with the check clock: it observes offline status when reports are older than 6 seconds or a disconnected state stream has cleared the active state, then accumulates the configured window. duration: 10 means an approximately 30-second offline observation window, not guaranteed delivery exactly 30 seconds after the last report. Notification network latency adds further delay.

Examples:

Offline notification:
- Name: Offline notification
- Description: Notify after approximately 30 seconds of continuous downtime.
- Rule:
  json
```
[{"type": "offline", "duration": 10}]
```
- Enabled: checked

CPU and memory monitoring:

Name: CPU+Memory

Rule:

json

[
  {"type": "cpu", "min": 0, "max": 50, "duration": 10},
  {"type": "memory", "min": 20, "max": 0, "duration": 20}
]

Enabled: checked

Send notifications for specific servers to specified notification groups:
- Scenario: Servers 1 and 2 send notifications to notification group A after being offline for 10 minutes; servers 3 and 4 send notifications to notification group B after being offline for 10 minutes.
- Rule 1 (notification group A):
  - Name: Servers 1 and 2 offline notification
  - Rule:
    json
```
[{"type": "offline", "duration": 200, "cover": 1, "ignore": {"1": true, "2": true}}]
```
  - Notification group: A
  - Enabled: checked
- Rule 2 (notification group B):
  - Name: Servers 3 and 4 offline notification
  - Rule:
    json
```
[{"type": "offline", "duration": 200, "cover": 1, "ignore": {"3": true, "4": true}}]
```
  - Notification group: B
  - Enabled: checked

Special Rule: Any-Period Traffic Notification

This can be used for monthly traffic monitoring.

type:
- transfer_in_cycle: inbound traffic during the cycle
- transfer_out_cycle: outbound traffic during the cycle
- transfer_all_cycle: total bidirectional traffic during the cycle
cycle_start: Statistics cycle start date in RFC3339 format, such as 2022-01-01T00:00:00+08:00.
cycle_interval: Number of statistics cycles, such as every 1 month.
cycle_unit: Statistics cycle unit. Optional values: hour, day, week, month, year.
min / max, cover, and ignore are the same as basic rules.

Example:

Monthly traffic over-limit notification:

Rule:

json

[
  {
    "type": "transfer_out_cycle",
    "max": 1099511627776,
    "cycle_start": "2022-01-01T00:00:00+08:00",
    "cycle_interval": 1,
    "cycle_unit": "month",
    "cover": 1,
    "ignore": {"3": true, "4": true}
  }
]

Description: A notification is triggered when monthly outbound traffic for servers 3 and 4 exceeds 1TB. The statistics cycle starts on the 1st day of each month.

TIP

If you are unsure about alert rules, use these third-party configuration generators to simplify setup. Nezha Monitor does not provide any guarantee for third-party generator functionality:

Nezha Rule Generator: suitable for most scenarios.
NEZHA Config Generator: focused on cycle traffic notification rule generation. Select “traffic monitoring code” at the top.

Notification Trigger Modes

Always: Continues attempting delivery during an incident. Successful sends are subject to notification mute intervals; not every report produces an actual delivery.
Once: Sends one successful notification for each incident/recovery transition. Failed sends are still retried rather than suppressing the notification merely because it was attempted.

Incident and recovery delivery for the same rule/server pair is serialized. A new state cancels retries for the old state that have not started, and recovery cannot overtake an in-flight incident request. Success is tracked per notification method in the group, so failed recipients remain retryable; an empty group is not a successful send. This does not guarantee exactly-once delivery at a third-party receiver.

Execute Tasks on Notification

If you need to execute a specific task while sending notifications, configure:

Incident tasks (fail_trigger_tasks): In Always mode, each evaluable failed check may trigger them; in Once mode, only a new incident triggers them.
Recovery tasks (recover_trigger_tasks): Trigger when a known incident returns to normal.

Configure tasks on the Tasks page first. Task execution is not conditional on successful notification delivery. Do not use Always mode for tasks that are unsafe to repeat.

Notification Settings ​

Flexible Notification Methods ​

Supported Placeholders ​

Bark Example ​

Slack Example (Contributed by @cantoblanco) ​

Get URL Parameters ​

ServerChan Example ​

Telegram Example (Contributed by @cantoblanco) ​

Get URL Parameters ​

wxpusher Example ​

Email Notification Example (Using SendCloud, Contributed by @cantoblanco, @erbanku) ​

Get URL Parameters ​

DingTalk Group Bot Example ​

Get URL Parameters ​

WeCom Group Bot Example (Contributed by @ChowRex) ​

Configuration Example ​

Feishu Group Bot Example (Contributed by @eya46) ​

Get URL Parameters ​

Matrix Notification Example ​

Parameter Description ​

Retrieve Access Token ​

Notification Configuration ​

Alert Rule Description ​

Basic Rules ​

Special Rule: Any-Period Traffic Notification ​

Notification Trigger Modes ​

Execute Tasks on Notification ​

Notification Settings

Flexible Notification Methods

Supported Placeholders

Bark Example

Slack Example (Contributed by @cantoblanco)

Get URL Parameters

ServerChan Example

Telegram Example (Contributed by @cantoblanco)

Get URL Parameters

wxpusher Example

Email Notification Example (Using SendCloud, Contributed by @cantoblanco, @erbanku)

Get URL Parameters

DingTalk Group Bot Example

Get URL Parameters

WeCom Group Bot Example (Contributed by @ChowRex)

Configuration Example

Feishu Group Bot Example (Contributed by @eya46)

Get URL Parameters

Matrix Notification Example

Parameter Description

Retrieve Access Token

Notification Configuration

Alert Rule Description

Basic Rules

Special Rule: Any-Period Traffic Notification

Notification Trigger Modes

Execute Tasks on Notification