http-service (Source)
The http-service source receives POST requests via HTTP and HTTPS
protocols in format such as text
, XML
and JSON
and sends responses
via its corresponding http-service-response sink correlated through a
unique source.id
. For request and response correlation, it generates a
messageId
upon each incoming request and expose it via transport
properties in the format trp:messageId
to correlate them with the
responses at the http-service-response sink. The request headers and
properties can be accessed via transport properties in the format
trp:<header>
. It also supports basic authentication to ensure events
are received from authorized users/systems.
Syntax
CREATE SOURCE <NAME> WITH (type="http-service", map.type="<STRING>", receiver.url="<STRING>", source.id="<STRING>", connection.timeout="<INT>", basic.auth.enabled="<STRING>", worker.count="<INT>", socket.idle.timeout="<INT>", ssl.verify.client="<STRING>", ssl.protocol="<STRING>", tls.store.type="<STRING>", ssl.configurations="<STRING>", request.size.validation.configurations="<STRING>", header.validation.configurations="<STRING>", server.bootstrap.configurations="<STRING>", trace.log.enabled="<BOOL>")
Query Parameters
Name | Description | Default Value | Possible Data Types | Optional | Dynamic |
---|---|---|---|---|---|
receiver.url | The URL on which events should be received. To enable SSL use https protocol in the url. | `http://0.0.0.0:9763//\` | STRING | Yes | No |
source.id | Identifier to correlate the http-service source to its corresponding http-service-response sinks to send responses. | STRING | No | No | |
connection.timeout | Connection timeout in millis. The system will send a timeout, if a corresponding response is not sent by an associated http-service-response sink within the given time. | 120000 | INT | Yes | No |
basic.auth.enabled | This only works in VM, Docker and Kubernetes. Where when enabled it authenticates each request using the Authorization:'Basic encodeBase64(username:Password)' header. | false | STRING | Yes | No |
worker.count | The number of active worker threads to serve the incoming events. By default the value is set to 1 to ensure events are processed in the same order they arrived. By increasing this value, higher performance can be achieved in the expense of loosing event ordering. | 1 | INT | Yes | No |
socket.idle.timeout | Idle timeout for HTTP connection in millis. | 120000 | INT | Yes | No |
ssl.verify.client | The type of client certificate verification. Supported values are require , optional . | - | STRING | Yes | No |
ssl.protocol | SSL/TLS protocol. | TLS | STRING | Yes | No |
tls.store.type | TLS store type. | JKS | STRING | Yes | No |
ssl.configurations | SSL/TSL configurations in format "'<key>:<value>','<key>:<value>'" . Some supported parameters: - SSL/TLS protocols: 'sslEnabledProtocols:TLSv1.1,TLSv1.2' - List of ciphers: 'ciphers:TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256' - Enable session creation: 'client.enable.session.creation:true' - Supported server names: 'server.suported.server.names:server' - Add HTTP SNIMatcher: 'server.supported.snimatchers:SNIMatcher' | - | STRING | Yes | No |
request.size.validation.configurations | Configurations to validate the HTTP request size. Expected format "'<key>:<value>','<key>:<value>'" . Some supported configurations : - Enable request size validation: 'request.size.validation:true' If request size is validated - Maximum request size: 'request.size.validation.maximum.value:2048' - Response status code when request size validation fails: 'request.size.validation.reject.status.code:401' - Response message when request size validation fails: 'request.size.validation.reject.message:Message is bigger than the valid size' - Response Content-Type when request size validation fails: 'request.size.validation.reject.message.content.type:plain/text' | - | STRING | Yes | No |
header.validation.configurations | Configurations to validate HTTP headers. Expected format "'<key>:<value>','<key>:<value>'" . Some supported configurations : - Enable header size validation: 'header.size.validation:true' If header size is validated - Maximum length of initial line: 'header.validation.maximum.request.line:4096' - Maximum length of all headers: 'header.validation.maximum.size:8192' - Maximum length of the content or each chunk: 'header.validation.maximum.chunk.size:8192' - Response status code when header validation fails: 'header.validation.reject.status.code:401' - Response message when header validation fails: 'header.validation.reject.message:Message header is bigger than the valid size' - Response Content-Type when header validation fails: 'header.validation.reject.message.content.type:plain/text' | - | STRING | Yes | No |
server.bootstrap.configurations | Server bootstrap configurations in format "'<key>:<value>','<key>:<value>'" . Some supported configurations : - Server connect timeout in millis: 'server.bootstrap.connect.timeout:15000' - Server socket timeout in seconds: 'server.bootstrap.socket.timeout:15' - Enable TCP no delay: 'server.bootstrap.nodelay:true' - Enable server keep alive: 'server.bootstrap.keepalive:true' - Send buffer size: 'server.bootstrap.sendbuffersize:1048576' - Receive buffer size: 'server.bootstrap.recievebuffersize:1048576' - Number of connections queued: 'server.bootstrap.socket.backlog:100' | - | STRING | Yes | No |
trace.log.enabled | Enable trace log for traffic monitoring. | false | BOOL | Yes | No |
System Parameters
Name | Description | Default Value | Possible Parameters |
---|---|---|---|
serverBootstrapBossGroupSize | Number of boss threads to accept incoming connections. | Number of available processors | Any positive integer |
serverBootstrapWorkerGroupSize | Number of worker threads to accept the connections from boss threads and perform non-blocking read and write from one or more channels. | (Number of available processors) * 2 | Any positive integer |
serverBootstrapClientGroupSize | Number of client threads to perform non-blocking read and write to one or more channels. | (Number of available processors) * 2 | Any positive integer |
defaultHost | The default host of the transport. | 0.0.0.0 | Any valid host |
defaultScheme | The default protocol. | http | http https |
defaultHttpPort | The default HTTP port when default scheme is http . | 8280 | Any valid port |
defaultHttpsPort | The default HTTPS port when default scheme is https . | 8243 | Any valid port |
keyStoreLocation | The default keystore file path. | `\${carbon.home}/resources/security/gdncarbon.jks` | Path to `.jks` file |
keyStorePassword | The default keystore password. | gdncarbon | Keystore password as string |
Example 1
CREATE SOURCE AddStream WITH (type='http-service', receiver.url='http://localhost:5005/add', source.id='adder', map.type='json, map.attributes="messageId='trp:messageId', value1='$.event.value1', value2='$.event.value2'") (messageId string, value1 long, value2 long);
CREATE SINK ResultStream WITH (type='http-service-response', source.id='adder', message.id='{{messageId}}', map.type='json') (messageId string, results long);
@info(name = 'query1')
insert into ResultStream
select messageId, value1 + value2 as results
from AddStream;
Above sample listens events on http://localhost:5005/stocks
url for
JSON messages on the format:
{
"event": {
"value1": 3,
"value2": 4
}
}
Map the vents into AddStream, process the events through query query1
,
and sends the results produced on ResultStream via http-service-response
sink on the message format:
{
"event": {
"results": 7
}
}