This REST API allows users of the IBM Event Streams service to administer Kafka topics associated with an instance of the service. You can use this API to perform the following operations:
- Create a Kafka topic
- List Kafka topics
- Get a Kafka topic
- Delete a Kafka topic
- Update a Kafka topic configuration
- List which topics are mirrored
- Replace selection of topics which are mirrored
- List active mirroring topics
The Admin REST API is also documented using swagger.
All requests support below authorization methods:
- Basic authorization with user and password. (
For both standard, enterprise and lite plans, user is 'token', password is the API key from
ibmcloud resource service-keys
for the service instance.) - Bearer authorization with bearer token. (This token can be either API key or JWT token obtained from IAM upon login to IBM Cloud. Use
ibmcloud iam oauth-tokens
to retrieve the token afteribmcloud login
) X-Auth-Token
header to be set to the API key. This header is deprecated.
Administration API endpoint is the kafka_admin_url
property in the service key for the service instance. This command can be used to retrieve this property.
$ ibmcloud resource service-key "${service_instance_key_name}" --output json | jq -r '.[]|.credentials.kafka_admin_url'
In the examples you must set and export environment variables as follows:
- Either the
API_KEY
orBEARER_TOKEN
to use for authentication. KAFKA_ADMIN_URL
to point to your Eventstreams admin endpoint.
In addition, the Content-type
header has to be set to application/json
.
Common HTTP status codes:
- 200: Request succeeded.
- 202: Request was accepted.
- 400: Invalid request JSON.
- 401: The authentication header is not set or provided information is not valid.
- 403: Not authorized to perform the operation. Usually it means the API key used is missing a certain role. More details on what role can perform what operation refers to this document.
- 404: Unable to find the topic with topic name given by user.
- 422: Semantically invalid request.
- 503: An error occurred handling the request.
Error responses carry a JSON body like the following:
{"error_code":50301,"message":"Unknown Kafka Error", "incident_id": "17afe715-0ff5-4c49-9acc-a4204244a331"}
Error codes are of the format HHHKK
where HHH
is the HTTP Status Code and KK
is the Kafka protocol error.
For E2E debugging purposes, the transaction ID of every request is returned in the HTTP header X-Global-Transaction-Id
.
If the header is set on the request, it will be honored. If not, it will be generated.
In the event of a non-200 error return code, the transaction ID is also returned in the JSON error response as incident_id
.
To run the example in examples/adminrest :-
Compile the code.
cd examples/adminrest && go build -o example
Or simply
make build
Set the required environment variables
# Set your API KEY (or a bearer token could be used by setting the BEARER_TOKEN environment variable instead, but not both)
export API_KEY="abc123456789"
# Set the Admin Endpoint to point to your cluster.
export KAFKA_ADMIN_URL="https://xyzclustername.svc01.region.eventstreams.test.cloud.ibm.com"
Run the example
./examples/adminrest/example
The following sections explain how the REST API works with examples.
// Code Setup
import (
"fmt"
"net/http"
"os"
"github.com/IBM/eventstreams-go-sdk/pkg/adminrestv1"
"github.com/IBM/go-sdk-core/v5/core"
)
// End Code Setup
Use one of the following methods to authenticate:
-
To authenticate using Basic Auth: Place these values into the Authorization header of the HTTP request in the form Basic where is the username and password joined by a single colon
:
base64 encoded. For example:echo -n "token:<APIKEY>" | base64
-
To authenticate using a bearer token: To obtain your token using the IBM Cloud CLI, first log in to IBM Cloud, then run the following command:
ibmcloud iam oauth-tokens
Place this token in the Authorization header of the HTTP request in the form Bearer. Both API key or JWT tokens are supported.
-
To authenticate directly using the api_key: Place the key directly as the value of the X-Auth-Token HTTP header.
Here's an example of how to create the authenticator using either an API key or a BEARER_TOKEN
// Create Authenticator
var authenticator core.Authenticator
if apiKey != "" {
var err error
// Create an Basic IAM authenticator.
authenticator, err = core.NewBasicAuthenticator("token", apiKey)
if err != nil {
fmt.Printf("failed to create new basic authenticator: %s\n", err.Error())
os.Exit(1)
}
} else {
var err error
// Create an IAM Bearer Token authenticator.
authenticator, err = core.NewBearerTokenAuthenticator(bearerToken)
if err != nil {
fmt.Printf("failed to create new bearer token authenticator: %s\n", err.Error())
os.Exit(1)
}
}
// End Authenticator
Create a new service object.
// Create Service
serviceAPI, serviceErr := adminrestv1.NewAdminrestV1(&adminrestv1.AdminrestV1Options{
URL: URL,
Authenticator: authenticator,
})
// End Create Service
To create a Kafka topic the admin REST SDK issues a POST request to the /admin/topics path. The body of the request contains a JSON document, for example:
{
"name": "topicname",
"partitions": 1,
"configs": {
"retentionMs": 86400000,
"cleanupPolicy": "delete"
}
}
The only required field is name. The partitions fields defaults to 1 if not set.
Expected HTTP status codes:
- 202: Topic creation request was accepted.
- 400: Invalid request JSON.
- 403: Not authorized to create topic.
- 422: Semantically invalid request.
If the request to create a Kafka topic succeeds then HTTP status code 202 (Accepted) is returned. If the operation fails then a HTTP status code of 422 (Un-processable Entity) is returned, and a JSON object containing additional information about the failure is returned as the body of the response.
func createTopic(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the createTopicOptionsModel.
createTopicOptionsModel := new(adminrestv1.CreateTopicOptions)
createTopicOptionsModel.Name = core.StringPtr("test-topic")
createTopicOptionsModel.PartitionCount = core.Int64Ptr(int64(1))
// Create the Topic.
response, operationErr := serviceAPI.CreateTopic(createTopicOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Creating Topics: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Creating Topic: status %d\n", response.StatusCode)
}
fmt.Printf("\tname: %s created\n", *createTopicOptionsModel.Name)
return nil
}
To delete a Kafka topic, the admin REST SDK issues a DELETE request to the /admin/topics/TOPICNAME
path (where TOPICNAME
is the name of the Kafka topic that you want to delete).
Expected return codes:
- 202: Topic deletion request was accepted.
- 403: Not authorized to delete topic.
- 404: Topic does not exist.
- 422: Semantically invalid request.
A 202 (Accepted) status code is returned if the REST API accepts the delete request or status code 422 (Un-processable Entity) if the delete request is rejected. If a delete request is rejected then the body of the HTTP response will contain a JSON object which provides additional information about why the request was rejected.
Kafka deletes topics asynchronously. Deleted topics may still appear in the response to a list topics request for a short period of time after the completion of a REST request to delete the topic.
func deleteTopic(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the DeleteTopicOptions model
deleteTopicOptionsModel := new(adminrestv1.DeleteTopicOptions)
deleteTopicOptionsModel.TopicName = core.StringPtr("test-topic")
// Delete Topic
response, operationErr := serviceAPI.DeleteTopic(deleteTopicOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Deleting Topic: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Deleting Topic: status %d\n", response.StatusCode)
}
fmt.Printf("\tname: %s deleted\n", *deleteTopicOptionsModel.TopicName)
return nil
}
You can list all of your Kafka topics by issuing a GET request to the
/admin/topics
path.
Expected status codes:
- 200: the topic list is returned as JSON in the following format:
[
{
"name": "topic1",
"partitions": 1,
"retentionMs": 86400000,
"cleanupPolicy": "delete"
},
{ "name": "topic2",
"partitions": 2,
"retentionMs": 86400000,
"cleanupPolicy": "delete"
}
]
A successful response will have HTTP status code 200 (OK) and contain an array of JSON objects, where each object represents a Kafka topic and has the following properties:
Property name | Description |
---|---|
name | The name of the Kafka topic. |
partitions | The number of partitions of the Kafka topic. |
retentionsMs | The retention period for messages on the topic (in ms). |
cleanupPolicy | The cleanup policy of the Kafka topic. |
func listTopics(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the ListTopicsOptions model
listTopicsOptionsModel := new(adminrestv1.ListTopicsOptions)
// Call ListTopics.
result, response, operationErr := serviceAPI.ListTopics(listTopicsOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Listing Topics" + operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusOK {
return fmt.Errorf("Error Listing Topics: status %d\n", response.StatusCode)
}
// Loop and print topics.
for _, topicDetail := range result {
fmt.Printf("\tname: %s\n", *topicDetail.Name)
}
return nil
}
To get a Kafka topic detail information, issue a GET request to the /admin/topics/TOPICNAME
path (where TOPICNAME
is the name of the Kafka topic that you want to get).
Expected status codes
- 200: Retrieve topic details successfully in following format:
{
"name": "MYTOPIC",
"partitions": 1,
"replicationFactor": 3,
"retentionMs": 86400000,
"cleanupPolicy": "delete",
"configs": {
"cleanup.policy": "delete",
"min.insync.replicas": "2",
"retention.bytes": "1073741824",
"retention.ms": "86400000",
"segment.bytes": "536870912"
},
"replicaAssignments": [
{
"id": 0,
"brokers": {
"replicas": [
3,
2,
4
]
}
}
]
}
- 403: Not authorized.
- 404: Topic does not exist.
func topicDetails(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the GetTopicOptions model
getTopicOptionsModel := new(adminrestv1.GetTopicOptions)
getTopicOptionsModel.TopicName = core.StringPtr("test-topic")
// Call List Topic Details.
result, response, operationErr := serviceAPI.GetTopic(getTopicOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Listing Topic Details" + operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusOK {
return fmt.Errorf("Error Listing Topic Details: status %d\n", response.StatusCode)
}
// Print topics details.
fmt.Printf("\tname: \t\t\t%s\n", *result.Name)
// The number of partitions.
fmt.Printf("\tno of partitions: \t%d\n", *result.Partitions)
// The number of replication factor.
fmt.Printf("\treplication factor: \t%d\n", *result.ReplicationFactor)
// // The value of config property 'retention.ms'.
fmt.Printf("\tretention (ms): \t%d\n", *result.RetentionMs)
// // The value of config property 'cleanup.policy'.
fmt.Printf("\tcleanup policy: \t%s\n", *result.CleanupPolicy)
// Configs *TopicConfigs
fmt.Printf("\ttopic configs: \t\t%+v\n", *result.Configs)
// The replia assignment of the topic.
// ReplicaAssignments []ReplicaAssignment
for _, assignment := range result.ReplicaAssignments {
fmt.Printf("\tassignment: \t\tid:%d, \tbrokers: %+v\n", assignment.ID, assignment.Brokers)
}
return nil
}
To increase a Kafka topic's partition number or to update a Kafka topic's configuration, issue a
PATCH
request to /admin/topics/TOPICNAME
with the following body:
(where TOPICNAME is the name of the Kafka topic that you want to update).
{
"new_total_partition_count": 4,
"configs": [
{
"name": "cleanup.policy",
"value": "compact"
}
]
}
Supported configuration keys are 'cleanup.policy', 'retention.ms', 'retention.bytes', 'segment.bytes', 'segment.ms', 'segment.index.bytes'. And partition number can only be increased, not decreased.
Expected status codes
- 202: Update topic request was accepted.
- 400: Invalid request JSON/number of partitions is invalid.
- 404: Topic specified does not exist.
- 422: Semantically invalid request.
func updateTopicDetails(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the UpdateTopicOptions model
updateTopicOptionsModel := new(adminrestv1.UpdateTopicOptions)
updateTopicOptionsModel.TopicName = core.StringPtr("test-topic")
updateTopicOptionsModel.NewTotalPartitionCount = core.Int64Ptr(int64(6))
// Invoke operation with valid options model.
response, operationErr := serviceAPI.UpdateTopic(updateTopicOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Updating Topic: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Updating Topics: status %d\n", response.StatusCode)
}
fmt.Printf("\tname: %s updated\n", *updateTopicOptionsModel.TopicName)
return nil
}
Mirroring user controls are only available on the target cluster in a mirroring environment.
To get the current topic selection, issue an GET request to /admin/mirroring/topic-selection
Expected status codes
- 200: Retrieved topic selection successfully in following format:
{
"includes": [
"^prefix1_.*",
"^prefix2_.*"
]
}
- 403: Unauthorized to use mirroring user controls.
- 404: Mirroring not enabled. The mirroring user control APIs are only available on the target cluster of a mirrored pair.
- 503: An error occurred handling the request.
func listMirroringTopicSelection(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the GetMirroringTopicSelectionOptions model
getMirroringTopicSelectionOptionsModel := new(adminrestv1.GetMirroringTopicSelectionOptions)
// Call GetMirroringTopicSelection.
result, response, operationErr := serviceAPI.GetMirroringTopicSelection(getMirroringTopicSelectionOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Listing Mirroring Topics: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Listing Mirroring Topics: status %d\n", response.StatusCode)
}
// Loop and print mirroring topics.
for _, topicName := range result.Includes {
fmt.Printf("\tname: %s\n", topicName)
}
return nil
}
Replace mirroring topic selection
Mirroring user controls are available on the target cluster in a mirroring environment.
To replace the current topic selection, issue a POST request to /admin/mirroring/topic-selection
Expected status codes
- 200: Replaced topic selection successfully. The new selection is returned in following format:
{
"includes": [
"^prefix1_.*",
"^prefix2_.*"
]
}
- 400: Invalid request. The request data cannot be parsed and used to replace the topic selection.
- 403: Unauthorized to use mirroring user controls.
- 404: Mirroring not enabled. The mirroring user control APIs are only available on the target cluster of a mirrored pair.
- 415: Unsupported media type. Content-Type header with application/json is required.
- 503: An error occurred handling the request.
func replaceMirroringTopicSelection(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the ReplaceMirroringTopicSelectionOptions model
replaceMirroringTopicSelectionOptionsModel := new(adminrestv1.ReplaceMirroringTopicSelectionOptions)
replaceMirroringTopicSelectionOptionsModel.Includes = []string{"test-topic"}
// Invoke operation with valid options model.
result, response, operationErr := serviceAPI.ReplaceMirroringTopicSelection(replaceMirroringTopicSelectionOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Replacing Mirroring Topics: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Replacing Mirroring Topics: status %d\n", response.StatusCode)
}
// Loop and print mirroring topics.
for _, topicName := range result.Includes {
fmt.Printf("\ttopic added: %s\n", topicName)
}
return nil
}
Mirroring user controls are available on the target cluster in a mirroring environment.
To get the list of currently mirrored topics, issue an GET request to /admin/mirroring/active-topics
Expected status codes
- 200: Retrieved active topics successfully in following format:
{
"active_topics": [
"topic1",
"topic2"
]
}
- 403: Unauthorized to use mirroring user controls.
- 404: Mirroring not enabled. The mirroring user control APIs are only available on the target cluster of a mirrored pair.
- 503: An error occurred handling the request.
func getMirroringActiveTopics(serviceAPI *adminrestv1.AdminrestV1) error {
// Construct an instance of the GetMirroringActiveTopicsOptions model
getMirroringActiveTopicsOptionsModel := new(adminrestv1.GetMirroringActiveTopicsOptions)
// Call GetMirroringActiveTopics.
result, response, operationErr := serviceAPI.GetMirroringActiveTopics(getMirroringActiveTopicsOptionsModel)
if operationErr != nil {
return fmt.Errorf("Error Listing Active Mirroring Topics: %s\n", operationErr.Error())
}
// Check the result.
if response.StatusCode != http.StatusAccepted {
return fmt.Errorf("Error Listing Active Mirroring Topics: status %d\n", response.StatusCode)
}
// Loop and print mirroring topics.
for _, topicName := range result.ActiveTopics {
fmt.Printf("\tname: %s\n", topicName)
}
return nil
}