Configuration

Configuration

Cluster

cluster.name

Description

Cluster name.

cluster.proto_dist

Description

Distributed Erlang cluster protocol type. Available values are:

• inet_tcp: using IPv4
• inet6_tcp: using IPv6
• inet_tls: using TLS, required to be used with node.ssl_dist_optfile configuration

cluster.discovery

Description

Cluster node discovery method. Available values are:

• manual: join the cluster manually
• static: Configure static nodes. Configure several fixed nodes, and the new node joins the cluster by connecting one of the fixed nodes.
• mcast: Use UDP multicast to discover nodes.
• dns: Use DNS A records to discover nodes.
• etcd: Use etcd to discover nodes.
• k8s: Use Kubernetes to discover nodes.

cluster.autoheal

Description

Enable or disable the automatic recovery mechanism for cluster network partitions.

cluster.autoclean

Description

Specify how long to delete offline nodes from the cluster.

cluster.static.seeds

Description

When using static clustering, specify a fixed list of nodes, separated by commas , between multiple nodes.

cluster.mcast.addr

Description

When using the mcast cluster, specify the multicast address.

cluster.mcast.ports

Description

When using the mcast cluster, specify the multicast port. If there are multiple ports, separate them with commas ,.

cluster.mcast.iface

Description

When using mcast cluster, specify which local IP address the node discovery service needs to bind to.

cluster.mcast.ttl

Description

When using mcast cluster, specify the Time-To-Live value of multicast.

cluster.mcast.loop

Description

When using mcast clustering, set whether multicast packets are delivered to the local loopback address.

cluster.dns.name

Description

When using the dns cluster, specify the name of the DNS A record. emqx will access the DNS A record to obtain a list of IP addresses, and then splice the APP name specified in cluster.dns.app to get a list of all nodes in the cluster.

Example

Set cluster.dns.app = emqx, and configure a DNS: mycluster.com, which points to 3 IP addresses:

192.168.0.100
192.168.0.101
192.168.0.102

Then get the list of cluster nodes as follows:

emqx@192.168.0.100
emqx@192.168.0.101
emqx@192.168.0.102

cluster.dns.app

Description

When using dns cluster, it is used to splice the IP list obtained from cluster.dns.name to get a list of node names.

cluster.etcd.server

Description

When using etcd cluster, specify the address of etcd service. If there are multiple services, use commas to separate them.

cluster.etcd.prefix

Description

When using etcd cluster, specify the prefix of etcd path. Each node creates a path in etcd:

v2/keys/<prefix>/<cluster.name>/<node.name>

cluster.etcd.node_ttl

Description

When using etcd cluster, specify the expiration time of the node path in etcd.

cluster.etcd.ssl.keyfile

Description

When using SSL to connect to etcd, specify the client’s private key file.

cluster.etcd.ssl.certfile

Description

When using SSL to connect to etcd, specify the SSL client certificate file.

cluster.etcd.ssl.cacertfile

Description

When using SSL to connect to etcd, specify the CA certificate file for SSL.

cluster.k8s.apiserver

Description

When using the k8s cluster, specify the Kubernetes API Server. If there are multiple Servers, separate them with commas ,.

cluster.k8s.service_name

Description

When using k8s cluster, specify the service name of EMQX Broker in Kubernetes.

cluster.k8s.address_type

Description

When using k8s cluster, address_type is used to obtain the host list from the response of the Kubernetes interface.

Example

Specifying cluster.k8s.address_type as ip, it will get the list of IP addresses of emqx services from the Kubernetes interface:

172.16.122.31
172.16.122.32
172.16.122.33

Then splice with the app name specified by cluster.k8s.app_name configuration to get a list of emqx nodes:

emqx@172.16.122.31
emqx@172.16.122.32
emqx@172.16.122.33

cluster.k8s.app_name

Description

When using k8s clustering, app_name is used to splice with the obtained Host list to get the node list.

cluster.k8s.suffix

Description

When using the k8s method and specifying cluster.k8s.address_type as the dns type, you can set the suffix of the emqx node name, and splice with cluster.k8s.namespace to get a list of node names.

cluster.k8s.namespace

Description

When using the k8s method and specifying cluster.k8s.address_type as the dns type, you can set the namespace of the emqx node name, and splice with cluster.k8s.suffix to get a list of node names.

Example

Setting cluster.k8s.address_type to dns, you will get the dns list of emqx service from the Kubernetes interface:

172-16-122-31
172-16-122-32
172-16-122-33

Then splice with cluster.k8s.app_name = emqx，cluster.k8s.suffix = pod.cluster.local，cluster.k8s.namespace = default to get a list of emqx node names in the form of dns:

emqx@172-16-122-31.default.pod.cluster.local
emqx@172-16-122-32.default.pod.cluster.local
emqx@172-16-122-33.default.pod.cluster.local

node.name

Description

The node name. The format is <name> @ <host>. Where <host> can be an IP address or FQDN. See http://erlang.org/doc/reference_manual/distributed.html (opens new window) for details

node.cookie

Description

The cookie value used by the distributed Erlang cluster.

node.data_dir

Description

The node’s data directory, which is used to store Mnesia data files.

node.heartbeat

Description

System tuning parameters. This configuration will override the -heart parameter in the vm.args file.

Enable or disable Erlang runtime detection mechanism, and restart automatically when the runtime terminates. Use with care to avoid restarting the monitored process when emqx is closed manually.

node.async_threads

Description

System tuning parameters. This configuration will override the +A parameter in the vm.args file.

Set the number of threads in the asynchronous thread pool in Erlang runtime, see http://erlang.org/doc/man/erl.html (opens new window) for details.

node.process_limit

Description

System tuning parameters. This configuration will override the +P parameter in the vm.args file.

Set the maximum number of processes allowed by Erlang, which will affect the number of connections that emqx nodes can process. See http://erlang.org/doc/man/erl.html (opens new window) for details.

node.max_ports

Description

System tuning parameters. This configuration will override the +Q parameter in the vm.args file.

Set the maximum number of ports allowed by Erlang. See http://erlang.org/doc/man/erl.html (opens new window) for details.

node.dist_buffer_size

Description

System tuning parameters. This configuration will override the +zdbbl parameter in the vm.args file.

Set the maximum cache size used by Erlang distributed communication. See http://erlang.org/doc/man/erl.html (opens new window) for details.

node.max_ets_tables

Description

System tuning parameters. This configuration will override the +e parameter in the vm.args file.

Set the maximum number of ETS tables allowed in Erlang runtime. See http://erlang.org/doc/man/erl.html (opens new window) for details.

node.global_gc_interval

Description

System tuning parameters, which set how often Erlang runs to force a global garbage collection.

node.fullsweep_after

Description

System tuning parameters. This configuration will override the -env ERL_FULLSWEEP_AFTER parameter in the vm.args file.

Set how many times the generational GC will run before Erlang runs a fullsweep GC. For details, see http://erlang.org/doc/man/erlang.html#spawn_opt-4 (opens new window).

node.crash_dump

Description

Set the storage path and file name of the Erlang crash_dump file.

node.ssl_dist_optfile

Description

This configuration will override the -ssl_dist_optfile parameter in the vm.args file.

If you use SSL to establish an emqx cluster, you need to specify the SSL distributed protocol configuration file. It needs to be used with cluster.proto_dist = inet_tls.

node.dist_net_ticktime

Description

System tuning parameters. This configuration will override the -kernel net_ticktime parameter in the vm.args file.

Specifying how long time when a node has been unresponsive, it is considered to be down and disconnected. For details, see http://www.erlang.org/doc/man/kernel_app.html#net_ticktime (opens new window).

node.dist_use_interface

Description

The default is to use 0.0.0.0 to specify all network-interfaces to listen to, or to specify the IP of the network-interface to listen to.

node.dist_listen_min

Description

Set a TCP port range together with node.dist_listen_max. This port ranget is used for distribution to distributed Erlang as a listening port for distributed channels. Note that if a firewall is set between nodes, this port range needs to be placed into the firewall’s whitelist.

node.dist_listen_max

Description

Set a TCP port range together with node.dist_listen_min. This port range is used for distribution to distributed Erlang as a listening port for distributed channels. Note that if a firewall is set up between nodes, this port rangeneeds to be put in The firewall’s whitelist.

rpc.mode

Description

RPC mode. Synchronous or asynchronous mode is optional.

rpc.async_batch_size

Description

The maximum number of batch messages sent in asynchronous mode. Note that this configuration does not work in synchronous mode.

rpc.port_discovery

Description

manual: discover ports by tcp_server_port and tcp_client_port. stateless: discover ports in a stateless manner. If node name is emqx<N>@127.0.0.1, where the <N> is an integer, then the listening port will be 5370 + <N>

Default is manual when started from docker (environment variable override from docker-entrypoint) otherwise stateless.

rpc.tcp_server_ip

Description

Set the listening network-interface used by RPC local service. Use 0.0.0.0 to listen to all network-interfaces. NOTE: this config only takes effect when rpc.port_discovery is set to manual

rpc.tcp_server_port

Description

Set the listening port used by RPC local service NOTE: this config only takes effect when rpc.port_discovery is set to manual

rpc.tcp_client_num

Description

Set the number of RPC communication channels initiated by this node to each remote node. Set to 1 to ensure the order of messages. Keep the default value (half the number of CPU cores) to improve RPC throughput.

rpc.driver

Description

Transport-layer protocol used for communication between the brokers.

rpc.default_client_driver

Description

Transport-layer protocol used for communication between the brokers (client side). This parameter should match the value of rpc.driver.

rpc.enable_ssl

Description

Listening port used by RPC local service. Note that this config only takes effect when rpc.port_discovery is set to manual and driver is set to ssl.

rpc.certfile

Description

Path to TLS certificate file used to validate identity of the cluster nodes. This configuration is mandatory when rpc.driver is set to ssl, otherwise it doesn’t take any effect.

rpc.cacertfile

Description

Path to certification authority TLS certificate file used to validate rpc.certfile. This configuration is mandatory when rpc.driver is set to ssl, otherwise it doesn’t take any effect.

rpc.keyfile

Description

Path to the private key file for the rpc.certfile. This configuration is mandatory when rpc.driver is set to ssl, otherwise it doesn’t take any effect. Note: contents of this file are secret, so it’s necessary to set permissions to 600.

rpc.connect_timeout

Description

Timeout for establishing an RPC connection. It means how long will it give up after trying if the remote node does not respond when establishing a connection, .

rpc.send_timeout

Description

Timeout for sending, which means how long to give up after sending the message.

rpc.authentication_timeout

Description

RPC authentication timeout. It means how long it will give up if the remote node does not respond, .

rpc.call_receive_timeout

Description

The timeout period of RPC synchronous mode. It means how long it will take before giving up if the RPC synchronous call fails to receive a reply.

rpc.socket_keepalive_idle

Description

It means how long after the last packet was sent, keepalive probe packets are sent.

rpc.socket_keepalive_interval

Description

The interval between keepalive detection messages.

rpc.socket_keepalive_count

Description

For how many times if the keepalive probe message fails to receive a reply, the RPC connection is considered lost.

rpc.socket_sndbuf

Description

TCP tuning parameters. TCP sending buffer size.

rpc.socket_recbuf

Description

TCP tuning parameters. TCP receiving buffer size.

rpc.socket_buffer

Description

TCP tuning parameters. Socket buffer size in user mode.

log.to

Description

Where to output the log. The optional values are:

• off: Disable logging completely
• file: Only output log to file
• console: Only output logs to standard output (emqx console)
• both: output log to file and standard output at the same time (emqx console)

log.level

Description

Global log level. This includes the primary log level and all log handlers. For details, see log level and log handlers.

log.dir

Description

Log file directory.

log.file

Description

The prefix of the log file. For example, if you use the default value (log.file = emqx.log), the log file name will be emqx.log.1, emqx.log.2, …

log.chars_limit

Description

Set the maximum length of a single log message. If this length is exceeded, the log message will be truncated. -1 means no limit.

log.max_depth

Description

Maximum depth for Erlang term log formatting and Erlang process message queue inspection. Set ‘unlimited’ (without quotes) to print Erlang terms without depth limit.

log.rotation.size

Description

Set the size of a single log file. If it exceeds this size, the log file will be rolled to create a new log file.

log.rotation.count

Description

Set the total number of log files. If this number is exceeded, the next log file will overwrite the first file.

log.<level>.file

Description

Set a separate log file for a certain log level.

Example

Separately output info and above logs to info.log.N file:

log.info.file = info.log

Output error and error logs separately to the error.log.N file

log.error.file = error.log

log.max_depth

Description

Max depth when printing large data blob to log. Exceeding parts will be logge as ‘…’.

log.single_line

Description

Print logs in a single line if set to true. If set to false, information like stacktraces in crash logs may span multiple lines.

log.formatter

Description

Choose log format. text for free text, and json for structured logging.

log.formatter.text.date.format

NOTE: This config is available start from EMQX Opensource 4.3.15, 4.4.4 and EMQX Enterprise 4.3.10, 4.4.4.

Description

The timestamp format for the text logger. Can be one of rfc3339 or a FORMAT string.

Supported specifiers in the FORMAT string are:

Examples:

## 2022-06-02T14:23:09.230000 +08:00
log.formatter.text.date.format = %Y-%m-%dT%H:%M:%S.%6N %:z

## To make the timestamp look the same as that of version 4.2.x (2022-06-02 14:24:36.124):
log.formatter.text.date.format = %Y-%m-%d %H:%M:%S.%3N

authacl

allow_anonymous

Description

Whether to allow anonymous users to log in to the system.

Tip

It is recommended to disable this option in the production environment.

acl_nomatch

Description

When the ACL is not hit, allow or deny the publish/subscribe operation.

acl_file

Description

The default path of ACL file.

enable_acl_cache

Description

Whether to enable ACL caching.

acl_cache_max_size

Description

Maximum cache number of ACL rule.

acl_cache_ttl

Description

Maximum cache time of ACL rule.

acl_deny_action

Description

What to do after the ACL check fails.

• ignore：No operation
• disconnect：disconnect.

acl_order

Description

When using multiple ACL backends, this config can be used to define their order. The default value none means no explicit ordering, in which case the order depends on the plugin (or module) start/restart order. Use comma to separate the names (or aliases), for example jwt,http means jwt authentication should be checked before http. Supported aliases are: internal (or file), http, jwt, ldap, mnesia, mongo (or mongodb), mysql, pgsql (or postgres), redis. It is not necessary to enumerate all the in-use ACL backends here, if only a part of the backends are listed here, the ones not listed will be ordered at the end. When using a third-party plugin, there is no alias support for it, so it has to be the specific callback module name, e.g. my_auth_plugin_module.

Tip

Unknown names are silently discarded.

auth_order

Description

When using multiple authentication backends, this config can be used to define their order. This config is similar to acl_order, only there is no internal authentication backend.

Tip

Unknown names are silently discarded.

alias_enrichment_module

Description

Tip

This is a very specific, advanced option that should not be used unless recommended by EMQ.

Introduced in 4.4.11.

Specify a module that defines the enrich_with_aliases/2 function. This function will be used to enrich the client/channel information with clientid and/or common name aliases (or other enrichments the module may implement).

special_auth_module

Description

Tip

This is a very specific, advanced option that should not be used unless recommended by EMQ.

Introduced in 4.4.11.

Specify a module that defines the check_authn/2 function. This function will be used in the client.authenticate hook as a way to implement custom authentication logic.

mqtt

flapping_detect_policy

Description

Specify the Flapping inspection strategy.

Format: <threshold>,<duration>,<banned>.

For example, 30, 1m, 5m, it means that if the client disconnects 30 times within 1 minute, then login is prohibited for the next 5 minutes

mqtt.max_packet_size

Description

The maximum allowed length of MQTT messages.

mqtt.max_clientid_len

Description

The maximum allowed length of Client ID string.

mqtt.max_topic_levels

Description

The maximum allowed level of topics for client subscription. 0 means no limit.

Warning

Too many topic levels may cause performance problems during subscription.

mqtt.max_qos_allowed

Description

The maximum allowed QoS level for client to publish.

mqtt.max_topic_alias

Description

The maximum allowed number of topic aliases. 0 means that topic aliases are not supported.

mqtt.retain_available

Description

Whether to support Retain message.

mqtt.wildcard_subscription

Description

Whether to support subscribing to wildcard topics.

mqtt.shared_subscription

Description

Whether to support shared subscriptions.

mqtt.ignore_loop_deliver

Description

Whether to ignore the message sent by itself. If it is ignored, it means that EMQX Broker will not deliver this message to the sender of the message.

mqtt.strict_mode

Description

Whether to enable the strict check mode. The strict check mode will check the correctness of the MQTT message in more detail.

zone.external.idle_timeout

Description

The daze time after the TCP connection is established. If no packets are received within this time, the connection will be shutdown.

zone.external.enable_acl

Description

Whether to enable ACL check.

zone.external.enable_ban

Description

Whether to enable blacklist.

zone.external.enable_stats

Description

Whether to enable client status statistics.

zone.external.acl_deny_action

| Type | Optional Value | Default | | —— | —————————— - | ———— | | enum | ignore, disconnect | ignore |

Description

What to do after the ACL check fails.

• ignore：No any operation.
• disconnect：disconnect.

zone.external.force_gc_policy

Description

When a certain number of messages, or bytes, are received, a garbage collection is forced.

Format: <Number> | <Bytes>.

For example, 16000|16MB means that when 16000 messages are received, or a byte of 16MB flows in, a garbage collection is forced.

zone.external.force_shutdown_policy

Description

When the process message queue length, or the memory bytes reaches a certain value, the process is forced to close.

The “message queue” here refers to the “message mailbox” of the Erlang process, not the “mqueue” of QoS 1 and QoS 2.

Format: <Number> | <Bytes>.

For example, 32000|32MB means that when the process accumulates 32000 messages, or the process occupies memory up to 32MB, the process is closed.

zone.external.max_packet_size

Description

The maximum allowed length of MQTT packet.

zone.external.max_clientid_len

Description

The maximum length of Client ID string.

zone.external.max_topic_levels

Description

The maximum allowed level of topics for client subscription. 0 means no limit.

Warning

Too many topic levels may cause performance problems during subscription.

zone.external.max_qos_allowed

Description

The maximum QoS level allowed for the client to publish.

zone.external.max_topic_alias

Description

The maximum number of topic aliases. 0 means that topic aliases are not supported.

zone.external.retain_available

Description

Whether to support Retain message.

zone.external.wildcard_subscription

Description

Whether to support subscribing to wildcard topics.

zone.external.shared_subscription

Description

Whether to support shared subscriptions.

zone.external.server_keepalive

Description

Keepalive time specified by the server, used for MQTT v5.0 CONNACK messages

zone.external.keepalive_backoff

Description

Keepalive backoff index. If no data packet is received from the client within the time of Keepalive * backoff * 2, it is considered that the client has heartbeat timeout.

zone.external.max_subscriptions

Description

The maximum number of topics that a single client is allowed to subscribe to. 0 means no limit.

zone.external.upgrade_qos

Description

Allow EMQX Broker to force the QoS level of the message upgrading to the subscribed QoS level when publishing the message.

zone.external.max_inflight

Description

Inflight window size: The inflight window is used to store unacknowledged QoS 1 and QoS 2 messages.

zone.external.retry_interval

Description

Message retransmission interval: EMQX Broker checks whether message retransmission is required at each interval.

zone.external.max_awaiting_rel

Description

The maximum receiving window for QoS 2 messages, which configures how many QoS 2 messages from the client can be processed by EMQX Broker simultaneously. 0 means no limit.

zone.external.await_rel_timeout

Description

Time for QoS 2 message processing timeout. If the QoS PUBREL message has not been received after the timeout, the message is dropped from the receiving window.

zone.external.session_expiry_interval

Description

The default timeout period of the session, which is mainly used for MQTT v3.1 and v3.1.1 protocols. In MQTT v5.0, this value is usually carried in the client’s connection message.

zone.external.max_mqueue_len

Description

The maximum length of the message queue. When the flight window is full, or the client is offline, the message will be stored in the queue. 0 means no limit.

zone.external.mqueue_priorities

Description

Queue message priority configuration:

• none：no prioritization.
• <Spec>：A message priority table, which configures the priority of messages under a certain topic. For example:
  • topic/1=10: indicates that the message priority of the topic topic/1 is 10.
  • topic/1=10,topic/2=8: indicates that the priority of two topics is configured, which are 10 and 8 respectively.
  • Among them, the higher the priority value, the higher the priority level.

When the length of the message queue is limited, low priority messages will be dropped first.

zone.external.mqueue_default_priority

Description

The default priority level of the message.

zone.external.mqueue_store_qos0

Description

Whether the message queue stores QoS 0 messages.

zone.external.enable_flapping_detect

Description

Whether to enable Flapping check.

zone.external.mountpoint

Description

After topic mount point is configured, all subscribed and published topics will be prefixed by EMQX Broker.

The available placeholders are:

• %c：Client ID.
• %u：Username.

For example, if the mount point is set to user/%c/. , when the client with client ID tom publishes the topic open message, the topic actually routed in EMQX Broker is user/tom/open.

zone.external.use_username_as_clientid

Description

Whether to use the client’s Username as its Client ID.

zone.external.ignore_loop_deliver

Description

Whether to ignore the message sent by yourself. If ignored, it means that EMQX Broker will not deliver this message to the sender of the message.

zone.external.strict_mode

Description

Whether to enable the strict check mode. The strict check mode will check the correctness of the MQTT message in more detail.

zone.internal.allow_anonymous

Description

Whether to allow anonymous users to log in to the system.

zone.internal.enable_stats

Description

Whether to enable client status statistics.

zone.internal.enable_acl

Description

Whether to enable ACL check.

zone.internal.acl_deny_action

Description

What to do after the ACL check fails.

• ignore：No operation.
• disconnect：Disconnect.

zone.internal.force_gc_policy

Description

When a certain number of messages, or bytes, are received, a garbage collection is forced.

Format: <Number> | <Bytes>.

For example, 16000|16MB means that when 16000 messages are received, or a byte of 16MB flows in, a garbage collection is forced.

zone.internal.wildcard_subscription

Description

Whether to support subscribing to wildcard topics.

zone.internal.shared_subscription

Description

Whether to support shared subscriptions.

zone.internal.max_subscriptions

Description

The maximum number of topics that a single client is allowed to subscribe to. 0 means no limit.

zone.internal.max_inflight

Description

Inflight window size: The flight window is used to store unanswered QoS 1 and QoS 2 messages.

zone.internal.max_awaiting_rel

Description

The maximum receiving window for QoS 2 messages, that configures how many QoS 2 messages from the client can be processed by EMQX Broker simultaneously. 0 means no limit.

zone.internal.max_mqueue_len

Description

The maximum length of the message queue. When the flight window is full, or the client is offline, the message will be stored in the queue. 0 means no limit.

zone.internal.mqueue_store_qos0

Description

Whether the message queue stores QoS 0 messages.

zone.internal.enable_flapping_detect

Description

Whether to enable Flapping check.

zone.internal.force_shutdown_policy

Description

When the process message queue length, or the memory bytes reaches a certain value, the process is forced to close.

The “message queue” here refers to the “message mailbox” of the Erlang process, not the “mqueue” of QoS 1 and QoS 2.

Format: <Number> | <Bytes>.

For example, 32000|32MB means that when the process accumulates 32000 messages, or the process occupies memory up to 32MB, the process is closed.

zone.internal.mountpoint

Description

After topic mount point is configured, all subscribed and published topics will be prefixed by EMQX Broker.

The available placeholders are:

• %c：Client ID.
• %u：Username.

For example, if the mount point is set to user/%c/. , when the client with client ID tom publishes the topic open message, the topic actually routed in EMQX Broker is user/tom/open.

zone.internal.ignore_loop_deliver

Description

Whether to ignore the message sent by itself. If ignored, it means that EMQX Broker will not deliver this message to the sender of the message.

zone.internal.strict_mode

Description

Whether to enable the strict check mode. The strict check mode will check the correctness of the MQTT message in more detail.

zone.internal.bypass_auth_plugins

Description

Whether to allow clients under this zone to bypass the authentication step of the authentication plugin.

listener.tcp.external

Description

Configure the listening address of the MQTT / TCP listener named external.

Example

1883: monitors IPv4 0.0.0.0: 1883. 127.0.0.1: 1883: monitor address is 1883 port on the 127.0.0.1 network card. :: 1: 1883: monitors the IPv6 address as 1883 port on the :: 1 network card.

listener.tcp.external.acceptors

Description

The size of the listener’s receiving pool.

listener.tcp.external.max_connections

Description

The maximum number of concurrent connections allowed by the listener.

listener.tcp.external.max_conn_rate

Description

The maximum access rate allowed by the listener. Unit: pcs / sec

listener.tcp.external.active_n

Description

The number of times the listener continues to receive TCP packets.

listener.tcp.external.zone

Description

The configuration zone to which the listener belongs.

listener.tcp.external.rate_limit

Description

The rate limit of the listener. The format is <limit>,<duration>.

Example

100KB,10s：Limit the number of incoming bytes within 10 seconds not to exceed 100 KB.

listener.tcp.external.access.1

Description

List of ACL rules of the listener. It is used to set the white/black list of the connection layer.

Example

allow all：Allow all TCP connections. allow 192.168.0.0/24：Allow TCP connection with network address 192.168.0.0/24.

At the same time, this configuration can configure multiple rules:

listener.tcp.external.access.1 = deny 192.168.0.1
listener.tcp.external.access.2 = allow all

It means that all TCP connections except 192.168.0.1 are allowed.

listener.tcp.external.proxy_protocol

Description

Whether the listener enables Proxy Protocol support.

If the EMQX cluster is deployed behind HAProxy or Nginx, and you need to get the client’s real source IP address and port, you need to enable this configuration.

Proxy Protcol : https://www.haproxy.com/blog/haproxy/proxy-protocol (opens new window).

listener.tcp.external.proxy_protocol_timeout

Description

Set the timeout for Proxy Protocol parsing. If no Proxy Protocol packet is received within this time, EMQX Broker will close its connection.