megaease/easeprobe repository overview

⁠EaseProbe

EaseProbe is a simple, standalone, and lightWeight tool that can do health/status checking, written in Go.

⁠Table of Contents

• EaseProbe⁠
  • 1. Overview⁠
    • 1.1 Probe⁠
    • 1.2 Notification⁠
    • 1.3 Report⁠
    • 1.4 Channel⁠
    • 1.5 Administration⁠
    • 1.6 Prometheus Metrics Exporter⁠
  • 2. Getting Started⁠
    • 2.1 Build⁠
    • 2.2 Configure⁠
    • 2.3 Run⁠
  • 3. Configuration⁠
    • 3.1 HTTP Probe Configuration⁠
    • 3.2 TCP Probe Configuration⁠
    • 3.3 Shell Command Probe Configuration⁠
    • 3.4 SSH Command Probe Configuration⁠
    • 3.5 TLS Probe Configuration⁠
    • 3.6 Host Resource Usage Probe Configuration⁠
    • 3.7 Native Client Probe Configuration⁠
    • 3.8 Notification Configuration⁠
    • 3.9 Global Setting Configuration⁠
  • 4. Benchmark⁠
  • 5. Contributing⁠
  • 6. Community⁠
  • 7. License⁠

⁠1. Overview

EaseProbe is designed to do three kinds of work - Probe, Notify, and Report.

⁠1.1 Probe

EaseProbe supports the following probing methods: HTTP, TCP, Shell Command, SSH Command, Host Resource Usage, and Native Client.

Each probe is identified by the method it supports (eg http), a unique name (across all probes in the configuration file) and the method specific parameters.

On application startup, the configured probes are scheduled for their initial fire up based on the following criteria:

• Less than or equal to 60 total probers exist: the delay between initial prober fire-up is 1 second
• More than 60 total probers exist: the startup is scheduled based on the following equation timeGap = DefaultProbeInterval / numProbes

• HTTP. Checking the HTTP status code, Support mTLS, HTTP Basic Auth, and can set the Request Header/Body. ( HTTP Probe Configuration⁠ )

  http:
    # Some of the Software support the HTTP Query
    - name: ElasticSearch
      url: http://elasticsearch.server:9200
    - name: Prometheus
      url: http://prometheus:9090/graph
  

  Copy

• TCP. Just simply check whether the TCP connection can be established or not. ( TCP Probe Configuration⁠ )

  tcp:
    - name: Kafka
      host: kafka.server:9093
  

  Copy

• Shell. Run a Shell command and check the result. ( Shell Command Probe Configuration⁠ )

  shell:
    # run redis-cli ping and check the "PONG"
    - name: Redis (Local)
      cmd: "redis-cli"
      args:
        - "-h"
        - "127.0.0.1"
        - "ping"
      env:
        # set the `REDISCLI_AUTH` environment variable for redis password
        - "REDISCLI_AUTH=abc123"
      # check the command output, if does not contain the PONG, mark the status down
      contain : "PONG"
  

  Copy

• SSH. Run a remote command via SSH and check the result. Support the bastion/jump server (SSH Command Probe Configuration⁠)

  ssh:
    servers:
      - name : ServerX
        host: [email protected]:22
        password: xxxxxxx
        key: /Users/user/.ssh/id_rsa
        passphrase: xxxxxxx
        cmd: "ps auxwe | grep easeprobe | grep -v grep"
        contain: easeprobe
  

  Copy

• TLS. Ping the remote endpoint, can probe for revoked or expired certificates ( TLS Probe Configuration⁠ )

  tls:
    - name: expired test
      host: expired.badssl.com:443
  

  Copy

• Host. Run an SSH command on a remote host and check the CPU, Memory, and Disk usage. ( Host Load Probe⁠ )

  host:
    servers:
      - name : server
        host: [email protected]:22
        key: /path/to/server.pem
        threshold:
          cpu: 0.80  # cpu usage  80%
          mem: 0.70  # memory usage 70%
          disk: 0.90  # disk usage 90%
  

  Copy

• Client. Currently, support the following native client. Support the mTLS. ( refer to: Native Client Probe Configuration⁠ )

  • MySQL. Connect to the MySQL server and run the SHOW STATUS SQL.
  • Redis. Connect to the Redis server and run the PING command.
  • Memcache. Connect to a Memcache server and run the version command or check based on key/value checks.
  • MongoDB. Connect to MongoDB server and just ping server.
  • Kafka. Connect to Kafka server and list all topics.
  • PostgreSQL. Connect to PostgreSQL server and run SELECT 1 SQL.
  • Zookeeper. Connect to Zookeeper server and run get / command.

  Most of the clients support the additional validity check of data pulled from the service (such as checking a redis or memcache key for specific values). Check the documentation of the corresponding client for details on how to enable.

  client:
    - name: MySQL Native Client (local)
      driver: "mysql"
      host: "localhost:3306"
      user: "root"
      password: "pass"
      data: # Optional - Data to check
        "database:table:column:id:100": "value"
      # mTLS - Optional
      ca: /path/to/file.ca
      cert: /path/to/file.crt
      key: /path/to/file.key
  

  Copy

⁠1.2 Notification

EaseProbe supports the following notifications:

• Slack. Using Webhook for notification
• Discord. Using Webhook for notification
• Telegram. Using Telegram Bot for notification
• Teams. Support the Microsoft Teams⁠ notification.
• Email. Support multiple email addresses.
• AWS SNS. Support AWS Simple Notification Service.
• WeChat Work. Support Enterprise WeChat Work notification.
• DingTalk. Support the DingTalk notification.
• Lark. Support the Lark(Feishu) notification.
• SMS. Support SMS notification with multiple SMS service providers - Twilio⁠, Vonage(Nexmo)⁠, YunPain⁠
• Log. Write the notification into a log file or syslog.
• Shell. Run a shell command to notify the result. (see example⁠)

# Notification Configuration
notify:
  log:
    - name: log file # local log file
      file: /var/log/easeprobe.log
    - name: Remote syslog # syslog (!!! Not For Windows !!!)
      file: syslog # <-- must be "syslog" keyword
      host: 127.0.0.1:514 # remote syslog server - optional
      network: udp #remote syslog network [tcp, udp] - optional
  slack:
    - name: "MegaEase#Alert"
      webhook: "https://hooks.slack.com/services/........../....../....../"
  discord:
    - name: "MegaEase#Alert"
      webhook: "https://discord.com/api/webhooks/...../....../"
  telegram:
    - name: "MegaEase Alert Group"
      token: 1234567890:ABCDEFGHIJKLMNOPQRSTUVWXYZ # Bot Token
      chat_id: -123456789 # Channel / Group ID
  email:
    - name: "DevOps Mailing List"
      server: smtp.email.example.com:465
      username: [email protected]
      password: ********
      to: "[email protected];[email protected]"
  aws_sns:
    - name: AWS SNS
      region: us-west-2
      arn: arn:aws:sns:us-west-2:298305261856:xxxxx
      endpoint: https://sns.us-west-2.amazonaws.com
      credential:
        id: AWSXXXXXXXID
        key: XXXXXXXX/YYYYYYY
  wecom:
    - name: "wecom alert service"
      webhook: "https://qyapi.weixin.qq.com/cgi-bin/webhook/send?key=589f9674-a2aa-xxxxxxxx-16bb6c43034a" # wecom robot webhook
  dingtalk:
    - name: "dingtalk alert service"
      webhook: "https://oapi.dingtalk.com/robot/send?access_token=xxxx"
      secret: "" # sign secret if set
  lark:
    - name: "lark alert service"
      webhook: "https://open.feishu.cn/open-apis/bot/v2/hook/d5366199-xxxx-xxxx-bd81-a57d1dd95de4"
  sms:
    - name: "sms alert service"
      provider: "yunpian"
      key: xxxxxxxxxxxx # yunpian apikey
      mobile: 123456789,987654321 # mobile phone number, multiple phone number joint by `,`
      sign: "xxxxxxxx" # need to register; usually brand name
  teams:
      - name: "teams alert service"
        webhook: "https://outlook.office365.com/webhook/a1269812-6d10-44b1-abc5-b84f93580ba0@9e7b80c7-d1eb-4b52-8582-76f921e416d9/IncomingWebhook/3fdd6767bae44ac58e5995547d66a4e4/f332c8d9-3397-4ac5-957b-b8e3fc465a8c" # see https://docs.microsoft.com/en-us/outlook/actionable-messages/send-via-connectors
  shell: # EaseProbe set the environment variables -
         # (see the example: resources/scripts/notify/notify.sh)
    - name: "shell alert service"
      cmd: "/bin/bash"
      args:
        - "-c"
        - "/path/to/script.sh"
      env: # set the env to the notification command
        - "EASEPROBE=1"

Check the Notification Configuration⁠ to see how to configure it.

⁠1.3 Report

• SLA Report Notify. EaseProbe would send the daily, weekly, or monthly SLA report.

  settings:
    # SLA Report schedule
    sla:
      #  daily, weekly (Sunday), monthly (Last Day), none
      schedule: "weekly"
      # UTC time, the format is 'hour:min:sec'
      time: "23:59"
  

  Copy

• SLA Live Report. You can query the SLA Live Report

  The EaseProbe would listen on the 0.0.0.0:8181 port by default. And you can access the Live SLA report by the following URL:

  • HTML: http://localhost:8181/
  • JSON: http://localhost:8181/api/v1/sla

You can use the following URL query options for both HTML and JSON:

• refresh: (HTML only) refresh the page every given seconds (ex, ?refresh=30s refreshes the page every 30 seconds)
• pg & sz: page number and page size (ex, ?pg=2&sz=10 shows the second page with 10 probers), default page size is 100
• name: filter the probers that contain the value of name (ex, ?name=probe1 list the probers which name containing probe1)
• kind: filter the probers with the kind (ex, ?kind=http list the probers with kind http)
• ep: filter the probers with the endpoint (ex, ?ep=example.com list the probers which endpoint containing example.com)
• msg: filter the probers with the message (ex, ?msg=example list the probers which message containing example)
• status: filter the probers with specific status, accepted values up or down (ex. ?status=up list only probers with status up).
• gte: filter the probers with SLA greater than or equal to the given percentage (ex. ?gte=50 filter only hosts with SLA percentage >= 50%)
• lte:filter the probers with SLA less than or equal to the given percentage (ex. ?lte=90 filter only hosts with SLA percentage <= 90% )

Refer to the Global Setting Configuration⁠ to see how to configure the access log.

• SLA Data Persistence. Save the SLA statistics data on the disk.

  The SLA data would be persisted in $CWD/data/data.yaml by default. If you want to configure the path, you can do it in the settings section.

  When EaseProbe starts, it looks for the location of data.yaml and if found, loads the file and removes any probes that are no longer present in the configuration file. Setting a value of "-" for data: disables SLA persistence (eg data: "-").

  settings:
    sla:
      # SLA data persistence file path.
      # The default location is `$CWD/data/data.yaml`
      data: /path/to/data/file.yaml
  

  Copy

For more information, please check the Global Setting Configuration⁠

⁠1.4 Channel

The Channel is used for connecting the Probers and the Notifiers. It can be configured for every Prober and Notifier.

This feature could help you group the Probers and Notifiers into a logical group.

For example:

http:
   - name: probe A
     channels : [ Dev_Channel, Manager_Channel ]
shell:
   - name: probe B
     channels: [ Ops_Channel ]
notify:
   - discord: Discord
     channels: [ Dev_Channel, Ops_Channel ]
   - email: Gmail
     channels: [ Mgmt_Channel ]

Then, we will have the following diagram

┌───────┐          ┌──────────────┐
│Probe B├─────────►│ Mgmt_Channel ├────┐
└───────┘          └──────────────┘    │
                                       │
                                       │
                   ┌─────────────┐     │   ┌─────────┐
            ┌─────►│ Dev_Channel ├─────▼───► Discord │
            │      └─────────────┘         └─────────┘
┌───────┐   │
│Probe A├───┤
└───────┘   │
            │      ┌────────────┐          ┌─────────┐
            └─────►│ QA_Channel ├──────────►  Gmail  │
                   └────────────┘          └─────────┘

⁠1.5 Administration

There are some administration configuration options:

PID file

The EaseProbe would create a PID file (default $CWD/easeprobe.pid) when it starts. it can be configured by:

settings:
  pid: /var/run/easeprobe.pid

• If the file already exists, EaseProbe would overwrite it.
• If the file cannot be written, EaseProbe would exit with an error.

If you want to disable the PID file, you can configure the pid file to "".

settings:
  pid: "" # EaseProbe won't create a PID file

Log file Rotation

There are two types of the log files: Application Log and HTTP Access Log.

Both application and HTTP access logs will be displayed on StdOut by default. Both can be be configured by the log: directive such as:

log:
  file: /path/to/log/file
  self_rotate: true # default: true

If self_rotate is true, EaseProbe would rotate the log automatically, and the following options are available:

  size: 10 # max size of log file. default: 10M
  age: 7 # max age days of log file. default: 7 days
  backups: 5 # max backup log files. default: 5
  compress: true # compress. default: true

If self_rotate is false, EaseProbe will not rotate the log, and the log file will have to be rotated by a 3rd-party tool (such as logrotate) or manually by the administrator.

mv /path/to/easeprobe.log /path/to/easeprobe.log.0
kill -HUP `cat /path/to/easeprobe.pid`

EaseProbe accepts the HUP signal to rotate the log.

⁠1.6 Prometheus Metrics Exporter

EaseProbe supports Prometheus metrics exporter. The Prometheus endpoint is http://localhost:8181/metrics by default.

Currently, All of the Probers support the following metrics:

• total: the total number of probes
• duration: Probe duration in milliseconds
• status: Probe status
• SLA: Probe SLA percentage

And for the different Probers, the following metrics are available:

• HTTP Probe

  • status_code: HTTP status code
  • content_len: HTTP content length
  • dns_duration: DNS duration in milliseconds
  • connect_duration: TCP connection duration in milliseconds
  • tls_duration: TLS handshake duration in milliseconds
  • send_duration: HTTP send duration in milliseconds
  • wait_duration: HTTP wait duration in milliseconds
  • transfer_duration: HTTP transfer duration in milliseconds
  • total_duration: HTTP total duration in milliseconds

• TLS Probe

  • earliest_cert_expiry: last TLS chain expiry in timestamp seconds
  • last_chain_expiry_timestamp_seconds: earliest TLS cert expiry in Unix time

• Shell & SSH Probe

  • exit_code: exit code of the command
  • output_len: length of the output

• Host Probe

  • cpu: CPU usage in percentage
  • memory: memory usage in percentage
  • disk: disk usage in percentage

The following snapshot is the Grafana panel for host CPU metrics

Refer to the Global Setting Configuration⁠ for further details on how to configure the HTTP server.

⁠2. Getting Started

You can get started with EaseProbe, by any of the following methods:

• Download the release for your platform from https://github.com/megaease/easeprobe/releases⁠
• Use the available EaseProbe docker image docker run -it megaease/easeprobe
• Build easeprobe from sources

⁠2.1 Build

Compiler Go 1.18+ (Generics Programming Support)

Use make to build and produce the easeprobe binary file. The executable is produced under the build/bin directory

$ make

⁠2.2 Configure

Read the Configuration Guide⁠ to learn how to configure EaseProbe.

Create the configuration file - $CWD/config.yaml.

The following is an example of simple configuration file to get started:

http: # http probes
  - name: EaseProbe Github
    url: https://github.com/megaease/easeprobe
notify:
  log:
    - name: log file # local log file
      file: /var/log/easeprobe.log
settings:
  probe:
    timeout: 30s # the time out for all probes
    interval: 1m # probe every minute for all probes

⁠2.3 Run

Running the following command for the local test

$ build/bin/easeprobe -f config.yaml

• -f configuration file or URL. Can also be achieved by setting the environment variable PROBE_CONFIG
• -d dry run. Can also be achieved by setting the environment variable PROBE_DRY

⁠3. Configuration

EaseProbe can be configured by supplying a YAML file or URL to fetch configuration settings from.

By default, EaseProbe will look for its config.yaml on the current folder. This behavior can be changed by supplying the -f parameter.

easeprobe -f path/to/config.yaml
easeprobe -f https://example.com/config

The following environment variables can be used to fine-tune the request to the configuration file

• HTTP_AUTHORIZATION
• HTTP_TIMEOUT

And the configuration file should be versioned, the version should be aligned with the EaseProbe binary version.

version: v1.5.0

The following example configurations illustrate the EaseProbe supported features.

Note: All probes have the following options:

• timeout - the maximum time to wait for the probe to complete. default: 30s.
• interval - the interval time to run the probe. default: 1m.

⁠3.1 HTTP Probe Configuration

# HTTP Probe Configuration

http:
  # A Website
  - name: MegaEase Website (Global)
    url: https://megaease.com

  # Some of the Software support the HTTP Query
  - name: ElasticSearch
    url: http://elasticsearch.server:9200
  - name: Eureka
    url: http://eureka.server:8761
  - name: Prometheus
    url: http://prometheus:9090/graph

  # Spring Boot Application with Actuator Heath API
  - name: EaseService-Governance
    url: http://easeservice-mgmt-governance:38012/actuator/health
  - name: EaseService-Control
    url: http://easeservice-mgmt-control:38013/actuator/health
  - name: EaseService-Mesh
    url: http://easeservice-mgmt-mesh:38013/actuator/health

  # A completed HTTP Probe configuration
  - name: Special Website
    url: https://megaease.cn
    # Proxy setting, support sock5, http, https, for example:
    #   proxy: http://proxy.server:8080
    #   proxy: socks5://localhost:1085
    #   proxy: https://user:[email protected]:443
    # Also support `HTTP_PROXY` & `HTTPS_PROXY` environment variables
    proxy: http://proxy.server:8080
    # Request Method
    method: GET
    # Request Header
    headers:
      User-Agent: Customized User-Agent # default: "MegaEase EaseProbe / v1.6.0"
      X-head-one: xxxxxx
      X-head-two: yyyyyy
      X-head-THREE: zzzzzzX-
    content_encoding: text/json
    # Request Body
    body: '{ "FirstName": "Mega", "LastName" : "Ease", "UserName" : "megaease", "Email" : "[email protected]"}'
    # HTTP Basic Auth
    username: username
    password: password
    # mTLS
    ca: /path/to/file.ca
    cert: /path/to/file.crt
    key: /path/to/file.key
    # TLS
    insecure: true # skip any security checks, useful for self-signed and expired certs. default: false
    # HTTP successful response code range, default is [0, 499].
    success_code:
      - [200,206] # the code >=200 and <= 206
      - [300,308] # the code >=300 and <= 308
    # Response Checking
    contain: "success" # response body must contain this string, if not the probe is considered failed.
    not_contain: "failure" # response body must NOT contain this string, if it does the probe is considered failed.
    regex: false # if true, the contain and not_contain will be treated as regular expression. default: false
    # configuration
    timeout: 10s # default is 30 seconds

⁠3.2 TCP Probe Configuration

# TCP Probe Configuration
tcp:
  - name: SSH Service
    host: example.com:22
    timeout: 10s # default is 30 seconds
    interval: 2m # default is 60 seconds
    proxy: socks5://proxy.server:1080 # Optional. Only support socks5.
                                      # Also support the `ALL_PROXY` environment.
  - name: Kafka
    host: kafka.server:9093

⁠3.3 Shell Command Probe Configuration

The shell command probe is used to execute a shell command and check the output.

The following example shows how to configure the shell command probe.

# Shell Probe Configuration
shell:
  # A proxy curl shell script
  - name: Google Service
    cmd: "./resources/probe/scripts/proxy.curl.sh"
    args:
      - "socks5://127.0.0.1:1085"
      - "www.google.com"

  # run redis-cli ping and check the "PONG"
  - name: Redis (Local)
    cmd: "redis-cli"
    args:
      - "-h"
      - "127.0.0.1"
      - "ping"
    clean_env: true # Do not pass the OS environment variables to the command
                    # default: false
    env:
      # set the `REDISCLI_AUTH` environment variable for redis password
      - "REDISCLI_AUTH=abc123"
    # check the command output, if does not contain the PONG, mark the status down
    contain : "PONG"
    not_contain: "failure" # response body must NOT contain this string, if it does the probe is considered failed.
    regex: false # if true, the `contain` and `not_contain` will be treated as regular expression. default: false

  # Run Zookeeper command `stat` to check the zookeeper status
  - name: Zookeeper (Local)
    cmd: "/bin/sh"
    args:
      - "-c"
      - "echo stat | nc 127.0.0.1 2181"
    contain: "Mode:"

⁠3.4 SSH Command Probe Configuration

SSH probe is similar to Shell probe.

• Support Password and Private key authentication.
• Support the Bastion host tunnel.

The host supports the following configuration

• example.com
• example.com:22
• [email protected]:22

The following are examples of SSH probe configuration.

# SSH Probe Configuration
ssh:
  # SSH bastion host configuration
  bastion:
    aws: # bastion host