Set up monitors

The heartbeat.monitors section of the heartbeat.yml config file specifies the list of monitors that Heartbeat uses to check your remote hosts to determine if they are available. Each monitor item begins with a dash (-) and specifies the type of monitor to use, the hosts to check, and other settings that control Heartbeat behavior.

The following example configures three monitors, icmp, tcp, and http, and demonstrates how to use TCP Echo and HTTP response verification:

heartbeat.monitors:
- type: icmp
  schedule: '*/5 * * * * * *'
  hosts: ["myhost"]
- type: tcp
  schedule: '@every 5s'
  hosts: ["myhost:7"]  # default TCP Echo Protocol
  check.send: "Check"
  check.receive: "Check"
- type: http
  schedule: '@every 5s'
  urls: ["http://localhost:80/service/status"]
  check.response.status: 200
heartbeat.scheduler:
  limit: 10

You can specify the following options in the heartbeat.monitors section of the heartbeat.yml config file. These options are the same for all monitors. Each monitor type has additional configuration options that are specific to that monitor type.

The type of monitor to run. One of:

The tcp and http monitor types both support SSL/TLS and some proxy settings.

The monitor name. This value appears in the exported fields under the monitor field as the job name and the type field as the job type.

A Boolean value that specifies whether the module is enabled. If the enabled option is missing from the configuration block, the module is enabled by default.

A cron-like expression that specifies the task schedule. For example:

The schedule option uses a cron-like syntax based on this cronexpr implementation, but adds the @every keyword.

A Boolean value that specifies whether to ping using the ipv4 protocol if hostnames are configured. The default is true.

A Boolean value that specifies whether to ping using the ipv6 protocol if hostnames are configured. The default is true.

If mode is any, the monitor pings only one IP address for a hostname. If mode is all, the monitor pings all resolvable IPs for a hostname. The mode: all setting is useful if you are using a DNS-load balancer and want to ping every IP address for the specified hostname. The default is any.

The total running time for each ping test. This is the total time allowed for testing the connection and exchanging data. The default is 16 seconds (16s).

If the timeout is exceeded, Heartbeat publishes a service-down event. If the value specified for timeout is greater than schedule, intermediate checks will not be executed by the scheduler.

Optional fields that you can specify to add additional information to the output. For example, you might add fields that you can use for filtering log data. Fields can be scalar values, arrays, dictionaries, or any nested combination of these. By default, the fields that you specify here will be grouped under a fields sub-dictionary in the output document. To store the custom fields as top-level fields, set the fields_under_root option to true. If a duplicate field is declared in the general configuration, then its value will be overwritten by the value declared here.

If this option is set to true, the custom fields are stored as top-level fields in the output document instead of being grouped under a fields sub-dictionary. If the custom field names conflict with other field names added by Heartbeat, then the custom fields overwrite the other fields.

A list of tags that will be sent with the monitor event. This setting is optional.

A list of processors to apply to the data generated by the monitor.

See Filter and Enhance the exported data for information about specifying processors in your config.

The JSON file to watch for additional monitor configurations. The JSON file can contain multiple objects, each of which specifies a different monitor config. Heartbeat checks this file periodically and starts a new monitor instance for each new JSON object added to the file. For example, imagine that you add 10 new entries to the JSON file, each for a different hostname. When Heartbeat picks up the changes in the file, it merges the original config (hearbeat.yml) plus the JSON objects, and starts a monitor for each new host that you’ve configured. If you delete an object from the JSON file and it doesn’t exist in the main config, Heartbeat stops the monitor instance running for that object.

Each monitor has a unique ID that’s based on parameters like protocol, host, and port. If two monitors have the same ID, Heartbeat uses the settings that are defined in the last JSON object of the merged config. This means that you can specify settings in the JSON file that overwrite the settings in the main config. In this way, the configuration that you specify for the monitor in the main Heartbeat config file acts like a default config that you can live-reconfigure by specifying additional configurations in the external JSON file.

Example configuration:

heartbeat.monitors:
- type: tcp
  schedule: '*/5 * * * * * *'
  hosts: ["myhost"]
  watch.poll_file:
    path: {path.config}/monitors/dynamic.json
    interval: 5s

To reconfigure the settings specified in the example config, you could define the following JSON objects in dynamic.json:

{"hosts": ["myhost:1234"], "schedule": "*/15 * * * * * *"} 
{"hosts": ["tls://otherhost:479"], "ssl.certificate_authorities": ["path/to/ca/file.pem"]} 

These options configure Heartbeat to use ICMP (v4 and v6) Echo Requests to check the configured hosts. These options are valid when the type is icmp.

A list of hosts to ping.

The duration to wait before emitting another ICMP Echo Request. The default is 1 second (1s).

These options configure Heartbeat to connect via TCP and optionally verify the endpoint by sending and/or receiving a custom payload. These options are valid when the type is tcp.

A list of hosts to ping. The entries in the list can be:

A list of ports to ping if the host specified in hosts does not contain a port number.

Example configuration:

- type: tcp
  schedule: '@every 5s'
  hosts: ["myhost"]
  ports: [80, 9200, 5044]

An optional payload string to send to the remote host and the expected answer. If no payload is specified, the endpoint is assumed to be available if the connection attempt was successful. If send is specified without receive, any response is accepted as OK. If receive is specified without send, no payload is sent, but the client expects to receive a payload in the form of a "hello message" or "banner" on connect.

Example configuration:

- type: tcp
  schedule: '@every 5s'
  hosts: ["myhost"]
  ports: [7]
  check.send: 'Hello World'
  check.receive: 'Hello World'

The URL of the SOCKS5 proxy to use when connecting to the server. The value must be a URL with a scheme of socks5://.

If the SOCKS5 proxy server requires client authentication, then a username and password can be embedded in the URL as shown in the example.

  proxy_url: socks5://user:password@socks5-proxy:2233

When using a proxy, hostnames are resolved on the proxy server instead of on the client. You can change this behavior by setting the proxy_use_local_resolver option.

A Boolean value that determines whether hostnames are resolved locally instead of being resolved on the proxy server. The default value is false, which means that name resolution occurs on the proxy server.

The TLS/SSL connection settings. If the monitor is configured to use SSL, it will attempt an SSL handshake. If check is not configured, the monitor will only check to see if it can establish an SSL/TLS connection. This check can fail either at TCP level or during certificate validation.

Example configuration:

- type: tcp
  schedule: '@every 5s'
  hosts: ["myhost"]
  ports: [80, 9200, 5044]
  ssl:
    certificate_authorities: ['/etc/ca.crt']
    supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]

These options configure Heartbeat to connect via HTTP and optionally verify that the host returns the expected response. These options are valid when the type is http.

A list of URLs to ping.

Example configuration:

- type: http
  schedule: '@every 5s'
  urls: ["http://myhost:80"]

The HTTP proxy URL. This setting is optional. If not set, the HTTP_PROXY environment variable is used.

The username for authenticating with the server. The credentials are passed with the request. This setting is optional.

You need to specify credentials when your check.response settings require it. For example, you can check for a 403 response (check.response.status: 403) without setting credentials.

The password for authenticating with the server. This setting is optional.

The TLS/SSL connection settings for use with the HTTPS endpoint. If you don’t specify settings, the system defaults are used.

Example configuration:

- type: http
  schedule: '@every 5s'
  urls: ["https://myhost:443"]
  ssl:
    certificate_authorities: ['/etc/ca.crt']
    supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]

An optional request to send to the remote host and the expected response.

Example configuration:

- type: http
  schedule: '@every 5s'
  urls: ["http://myhost:80"]
  check.request.method: HEAD
  check.response.status: 200

Under check.request, specify these options:

Under check.response, specify these options:

The following configuration shows how to check the response when the body contains JSON:

- type: http
  schedule: '@every 5s'
  urls: ["https://myhost:80"]
check.request:
  method: GET
  headers:
    'X-API-Key': '12345-mykey-67890'
check.response:
  status: 200
  body: '{"status": "ok"}'

You specify options under heartbeat.scheduler to control the behavior of the task scheduler.

Example configuration:

heartbeat.scheduler:
  limit: 10
  location: 'UTC-08:00'

In the example, setting limit to 10 guarantees that only 10 concurrent I/O tasks will be active. An I/O task can be the actual check or resolving an address via DNS.

The number of concurrent I/O tasks that Heartbeat is allowed to execute. If set to 0, there is no limit. The default is 0.

Most operating systems set a file descriptor limit of 1024. For Heartbeat to operate correctly and not accidentally block libbeat output, the value that you specify for limit should be below the configured ulimit.

The timezone for the scheduler. By default the scheduler uses localtime.