REST API requests must be authenticated. This authentication is only applicable for interacting with the Kafka REST API directly. You do not need the token to access the Kafka nodes themselves.
If you are using Enterprise DC/OS, follow these instructions to create a service account and an authentication token. You can then configure your service to automatically refresh the authentication token when it expires. To get started more quickly, you can also get the authentication token without a service account, but you will need to manually refresh the token.
If you are using open source DC/OS, follow these instructions to pass your authentication token to the DC/OS endpoint.
Once you have the authentication token, you can store it in an environment variable and reference it in your REST API calls:
$ export auth_token=uSeR_t0k3n
The curl examples in this document assume that an auth token has been stored in an environment variable named auth_token.
If you are using Enterprise DC/OS, the security mode of your installation may also require the --ca-cert flag when making REST calls. Refer to Obtaining and passing the DC/OS certificate in cURL requests for information on how to use the --cacert flag. If your security mode is disabled, do not use the --ca-cert flag.
For ongoing maintenance of the Kafka cluster itself, the Kafka service exposes an HTTP API whose structure is designed to roughly match the tools provided by the Kafka distribution, such as bin/kafka-topics.sh.
The examples here provide equivalent commands using both the DC/OS CLI (with the kafka CLI module installed) and curl. These examples assume a service named kafka (the default), and the curl examples assume a DC/OS cluster path of <dcos_url>. Replace these with appropriate values as needed.
The dcos kafka CLI commands have a --name argument, allowing the user to specify which Kafka instance to query. The value defaults to kafka, so it's technically redundant to specify --name=kafka in these examples.
Kafka comes with many useful tools of its own that often require either Zookeeper connection information or the list of broker endpoints. This information can be retrieved in an easily consumable format from the /connection endpoint:
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/endpoints/broker"
GET /service/kafka/v1/connection HTTP/1.1
{
"vips": [
"broker.kafka.l4lb.thisdcos.directory:9092"
],
"address": [
"10.0.0.35:1028",
"10.0.1.249:1030"
],
"dns": [
"kafka-0-broker.kafka.autoip.dcos.thisdcos.directory:1028",
"kafka-1-broker.kafka.autoip.dcos.thisdcos.directory:1030"
],
}
The same information can be retrieved through the DC/OS CLI:
$ dcos kafka endpoints broker
{
"vips": [
"broker.kafka.l4lb.thisdcos.directory:9092"
],
"address": [
"10.0.0.35:1028",
"10.0.1.249:1030"
],
"dns": [
"kafka-0-broker.kafka.autoip.dcos.thisdcos.directory:1028",
"kafka-1-broker.kafka.autoip.dcos.thisdcos.directory:1030"
],
}
Increase the BROKER_COUNT value via Marathon. This should be rolled as in any other configuration update.
$ dcos kafka --name=kafka broker list
{
"brokers": [
"0",
"1",
"2"
]
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/brokers"
GET /service/kafka/v1/brokers HTTP/1.1
{
"brokers": [
"0",
"1",
"2"
]
}
Restarts the broker in-place.
$ dcos kafka --name=kafka broker restart 0
[
"broker-0__9c426c50-1087-475c-aa36-cd00d24ccebb"
]
$ curl -X PUT -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/brokers/0"
PUT /service/kafka/v1/brokers/0 HTTP/1.1
[
"broker-0__9c426c50-1087-475c-aa36-cd00d24ccebb"
]
Restarts the broker and replaces its existing resource/volume allocations. The new broker instance may also be placed on a different machine.
$ dcos kafka --name=kafka broker replace 0
[
"broker-0__9c426c50-1087-475c-aa36-cd00d24ccebb"
]
$ curl -X PUT -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/brokers/0?replace=true"
PUT /service/kafka/v1/brokers/0 HTTP/1.1
[
"broker-0__9c426c50-1087-475c-aa36-cd00d24ccebb"
]
These operations mirror what is available with bin/kafka-topics.sh.
$ dcos kafka --name=kafka topic list
[
"topic1",
"topic0"
]
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics"
GET /service/kafka/v1/topics HTTP/1.1
[
"topic1",
"topic0"
]
$ dcos kafka --name=kafka topic describe topic1
{
"partitions": [
{
"0": {
"controller_epoch": 1,
"isr": [
0,
1,
2
],
"leader": 0,
"leader_epoch": 0,
"version": 1
}
},
{
"1": {
"controller_epoch": 1,
"isr": [
1,
2,
0
],
"leader": 1,
"leader_epoch": 0,
"version": 1
}
},
{
"2": {
"controller_epoch": 1,
"isr": [
2,
0,
1
],
"leader": 2,
"leader_epoch": 0,
"version": 1
}
}
]
}
$ curl -X POST -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/topic1"
GET /service/kafka/v1/topics/topic1 HTTP/1.1
{
"partitions": [
{
"0": {
"controller_epoch": 1,
"isr": [
0,
1,
2
],
"leader": 0,
"leader_epoch": 0,
"version": 1
}
},
{
"1": {
"controller_epoch": 1,
"isr": [
1,
2,
0
],
"leader": 1,
"leader_epoch": 0,
"version": 1
}
},
{
"2": {
"controller_epoch": 1,
"isr": [
2,
0,
1
],
"leader": 2,
"leader_epoch": 0,
"version": 1
}
}
]
}
$ dcos kafka --name=kafka topic create topic1 --partitions=3 --replication=3
{
"message": "Output: Created topic "topic1".n"
}
$ curl -X POST -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics?name=topic1&partitions=3&replication=3"
POST /service/kafka/v1/topics?replication=3&name=topic1&partitions=3 HTTP/1.1
{
"message": "Output: Created topic "topic1".n"
}
There is an optional --time parameter which may be set to either "first", "last", or a timestamp in milliseconds as described in the Kafka documentation.
$ dcos kafka --name=kafka topic offsets topic1 --time=last
[
{
"2": "334"
},
{
"1": "333"
},
{
"0": "333"
}
]
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/topic1/offsets?time=last"
GET /service/kafka/v1/topics/topic1/offsets?time=last HTTP/1.1
[
{
"2": "334"
},
{
"1": "333"
},
{
"0": "333"
}
]
$ dcos kafka --name=kafka topic partitions topic1 2
{
"message": "Output: WARNING: If partitions are increased for a topic that has a key, the partition logic or ordering of the messages will be affectednAdding partitions succeeded!n"
}
$ curl -X PUT -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/topic1?operation=partitions&partitions=2"
PUT /service/kafka/v1/topics/topic1?operation=partitions&partitions=2 HTTP/1.1
{
"message": "Output: WARNING: If partitions are increased for a topic that has a key, the partition logic or ordering of the messages will be affectednAdding partitions succeeded!n"
}
$ dcos kafka --name=kafka topic producer_test topic1 10
{
"message": "10 records sent, 70.422535 records/sec (0.07 MB/sec), 24.20 ms avg latency, 133.00 ms max latency, 13 ms 50th, 133 ms 95th, 133 ms 99th, 133 ms 99.9th.n"
}
$ curl -X PUT -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/topic1?operation=producer-test&messages=10"
PUT /service/kafka/v1/topics/topic1?operation=producer-test&messages=10 HTTP/1.1
{
"message": "10 records sent, 70.422535 records/sec (0.07 MB/sec), 24.20 ms avg latency, 133.00 ms max latency, 13 ms 50th, 133 ms 95th, 133 ms 99th, 133 ms 99.9th.n"
}
Runs the equivalent of the following command from the machine running the Kafka Scheduler:
kafka-producer-perf-test.sh
--topic <topic>
--num-records <messages>
--throughput 100000
--record-size 1024
--producer-props bootstrap.servers=<current broker endpoints>
$ dcos kafka --name=kafka topic delete topic1
{
"message": "Topic topic1 is marked for deletion.nNote: This will have no impact if delete.topic.enable is not set to true.n"
}
$ curl -X DELETE -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/topic1"
DELETE /service/kafka/v1/topics/topic1 HTTP/1.1
{
"message": "Topic topic1 is marked for deletion.nNote: This will have no impact if delete.topic.enable is not set to true.n"
}
Note the warning in the output from the commands above. You can change the indicated "delete.topic.enable" configuration value as a configuration change.
$ dcos kafka --name=kafka topic under_replicated_partitions
{
"message": ""
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/under_replicated_partitions"
GET /service/kafka/v1/topics/under_replicated_partitions HTTP/1.1
{
"message": ""
}
$ dcos kafka --name=kafka topic unavailable_partitions
{
"message": ""
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/topics/unavailable_partitions"
GET /service/kafka/v1/topics/unavailable_partitions HTTP/1.1
{
"message": ""
}
Send a GET request to the /v1/state/properties/suppressed endpoint to learn if Kafka is in a suppressed state and not receiving offers. If a service does not need offers, Mesos can "suppress" it so that other services are not starved for resources.
You can use this request to troubleshoot: if you think Kafka should be receiving resource offers, but is not, you can use this API call to see if Kafka is suppressed. You will receive a response of true or false.
curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/state/properties/suppressed"
These operations relate to viewing the service's configuration history.
$ dcos kafka --name=kafka config list
[
"319ebe89-42e2-40e2-9169-8568e2421023",
"294235f2-8504-4194-b43d-664443f2132b"
]
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/configurations"
GET /service/kafka/v1/configurations HTTP/1.1
[
"319ebe89-42e2-40e2-9169-8568e2421023",
"294235f2-8504-4194-b43d-664443f2132b"
]
This configuration shows a default per-broker memory allocation of 2048 (configured via the BROKER_MEM parameter):
$ dcos kafka --name=kafka config describe 319ebe89-42e2-40e2-9169-8568e2421023
{
"brokerConfiguration": {
"containerHookUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/container-hook-0.2.5.tgz",
"cpus": 1,
"disk": 5000,
"diskType": "ROOT",
"javaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/jre-8u72-linux-x64.tar.gz",
"kafkaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/kafka_2.10-0.9.0.1.tgz",
"mem": 2048, // <<--
"overriderUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/overrider.zip"
},
"kafkaConfiguration": {
[...]
},
"serviceConfiguration": {
"count": 3,
"name": "kafka",
"phaseStrategy": "INSTALL",
"placementStrategy": "NODE",
"principal": "kafka-principal",
"role": "kafka-role",
"user": ""
}
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/configurations/319ebe89-42e2-40e2-9169-8568e2421023"
GET /service/kafka/v1/configurations/319ebe89-42e2-40e2-9169-8568e2421023 HTTP/1.1
{
"brokerConfiguration": {
"containerHookUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/container-hook-0.2.5.tgz",
"cpus": 1,
"disk": 5000,
"diskType": "ROOT",
"javaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/jre-8u72-linux-x64.tar.gz",
"kafkaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/kafka_2.10-0.9.0.1.tgz",
"mem": 2048, // <<--
"overriderUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/overrider.zip"
},
"kafkaConfiguration": {
[...]
},
"serviceConfiguration": {
"count": 3,
"name": "kafka",
"phaseStrategy": "INSTALL",
"placementStrategy": "NODE",
"principal": "kafka-principal",
"role": "kafka-role",
"user": ""
}
}
The target configuration, meanwhile, shows an increase of configured per-broker memory from 2048 to 4096 (again, configured as BROKER_MEM):
$ dcos kafka --name=kafka config target
{
"brokerConfiguration": {
"containerHookUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/container-hook-0.2.5.tgz",
"cpus": 1,
"disk": 5000,
"diskType": "ROOT",
"javaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/jre-8u72-linux-x64.tar.gz",
"kafkaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/kafka_2.10-0.9.0.1.tgz",
"mem": 4096, // <<--
"overriderUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/overrider.zip"
},
"kafkaConfiguration": {
[...]
},
"serviceConfiguration": {
"count": 3,
"name": "kafka",
"phaseStrategy": "INSTALL",
"placementStrategy": "NODE",
"principal": "kafka-principal",
"role": "kafka-role",
"user": ""
}
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/configurations/target"
GET /service/kafka/v1/configurations/target HTTP/1.1
{
"brokerConfiguration": {
"containerHookUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/container-hook-0.2.5.tgz",
"cpus": 1,
"disk": 5000,
"diskType": "ROOT",
"javaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/jre-8u72-linux-x64.tar.gz",
"kafkaUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/kafka_2.10-0.9.0.1.tgz",
"mem": 4096, // <<--
"overriderUri": "https://s3.amazonaws.com/downloads.mesosphere.io/kafka/assets/0.2.5-0.9.0.1/overrider.zip"
},
"kafkaConfiguration": {
[...]
},
"serviceConfiguration": {
"count": 3,
"name": "kafka",
"phaseStrategy": "INSTALL",
"placementStrategy": "NODE",
"principal": "kafka-principal",
"role": "kafka-role",
"user": ""
}
}
These options relate to viewing and controlling rollouts and configuration updates.
Displays all Phases and Steps in the service Plan. If a rollout is currently in progress, this returns a 503 HTTP code with response content otherwise unchanged.
$ dcos kafka --name=kafka plan
GET /service/kafka/v1/plan HTTP/1.1
{
"phases": [
{
"id": "1915bcad-1235-400f-8406-4ac7555a7d34",
"name": "Reconciliation",
"steps": [
{
"id": "9854a67d-7803-46d0-b278-402785fe3199",
"status": "COMPLETE",
"name": "Reconciliation",
"message": "Reconciliation complete"
}
],
"status": "COMPLETE"
},
{
"id": "3e72c258-1ead-465f-871e-2a305d29124c",
"name": "Update to: 329ef254-7331-48dc-a476-8a0e45752871",
"steps": [
{
"id": "ebf4cb02-1011-452a-897a-8c4083188bb2",
"status": "COMPLETE",
"name": "broker-0",
"message": "Broker-0 is COMPLETE"
},
{
"id": "ff9e74a7-04fd-45b7-b44c-00467aaacd5b",
"status": "IN_PROGRESS",
"name": "broker-1",
"message": "Broker-1 is IN_PROGRESS"
},
{
"id": "a2ba3969-cb18-4a05-abd0-4186afe0f840",
"status": "PENDING",
"name": "broker-2",
"message": "Broker-2 is PENDING"
}
],
"status": "IN_PROGRESS"
}
],
"errors": [],
"status": "IN_PROGRESS"
}
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/plan"
GET /service/kafka/v1/plan HTTP/1.1
{
"phases": [
{
"id": "1915bcad-1235-400f-8406-4ac7555a7d34",
"name": "Reconciliation",
"steps": [
{
"id": "9854a67d-7803-46d0-b278-402785fe3199",
"status": "COMPLETE",
"name": "Reconciliation",
"message": "Reconciliation complete"
}
],
"status": "COMPLETE"
},
{
"id": "3e72c258-1ead-465f-871e-2a305d29124c",
"name": "Update to: 329ef254-7331-48dc-a476-8a0e45752871",
"steps": [
{
"id": "ebf4cb02-1011-452a-897a-8c4083188bb2",
"status": "COMPLETE",
"name": "broker-0",
"message": "Broker-0 is COMPLETE"
},
{
"id": "ff9e74a7-04fd-45b7-b44c-00467aaacd5b",
"status": "COMPLETE",
"name": "broker-1",
"message": "Broker-1 is COMPLETE"
},
{
"id": "a2ba3969-cb18-4a05-abd0-4186afe0f840",
"status": "COMPLETE",
"name": "broker-2",
"message": "Broker-2 is COMPLETE"
}
],
"status": "COMPLETE"
}
],
"errors": [],
"status": "COMPLETE"
}
These operations are only applicable when PHASE_STRATEGY is set to STAGE, they have no effect when it is set to INSTALL. See the Changing Configuration at Runtime part of the Configuring section for more information.
$ dcos kafka --name=kafka continue
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/plan/continue"
$ dcos kafka --name=kafka interrupt
$ curl -H "Authorization: token=$auth_token" "<dcos_url>/service/kafka/v1/plan/interrupt"