blob: fee9e935ae1d5c790c27beb242dd8a30451a13b3 [file] [log] [blame]
<!DOCTYPE html>
<html lang="en">
<!--
Licensed to the Apache Software Foundation (ASF) under one or more
contributor license agreements. See the NOTICE file distributed with
this work for additional information regarding copyright ownership.
The ASF licenses this file to You under the Apache License, Version 2.0
(the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
<head>
<meta charset="utf-8" />
<title>SiteToSiteProvenanceReportingTask</title>
<link rel="stylesheet" href="../../../../../css/component-usage.css" type="text/css" />
</head>
<body>
<p>
The Site-to-Site Provenance Reporting Task allows the user to publish all of the Provenance Events from a NiFi instance back to
the same NiFi instance or another NiFi instance. This provides a great deal of power because it allows the user to make use of
all of the different Processors that are available in NiFi in order to process or distribute that data. When possible, it is
advisable to send the Provenance data to a different NiFi instance than the one that this Reporting Task is running on, because
when the data is received over Site-to-Site and processed, that in and of itself will generate Provenance events. As a result, there
is a cycle that is created. However, the data is sent in batches (1,000 by default). This means that for each batch of Provenance events
that are sent back to NiFi, the receiving NiFi will have to generate only a single event per component.
</p>
<p>
By default, when published to a NiFi instance, the Provenance data is sent as a JSON array. However, the user can define a Record Writer
and directly specify the output format and data with the assumption that the input schema is defined as follows:
</p>
<pre>
<code>
{
"type" : "record",
"name" : "provenance",
"namespace" : "provenance",
"fields": [
{ "name": "eventId", "type": "string" },
{ "name": "eventOrdinal", "type": "long" },
{ "name": "eventType", "type": "string" },
{ "name": "timestampMillis", "type": "long" },
{ "name": "durationMillis", "type": "long" },
{ "name": "lineageStart", "type": { "type": "long", "logicalType": "timestamp-millis" } },
{ "name": "details", "type": ["null", "string"] },
{ "name": "componentId", "type": ["null", "string"] },
{ "name": "componentType", "type": ["null", "string"] },
{ "name": "componentName", "type": ["null", "string"] },
{ "name": "processGroupId", "type": ["null", "string"] },
{ "name": "processGroupName", "type": ["null", "string"] },
{ "name": "entityId", "type": ["null", "string"] },
{ "name": "entityType", "type": ["null", "string"] },
{ "name": "entitySize", "type": ["null", "long"] },
{ "name": "previousEntitySize", "type": ["null", "long"] },
{ "name": "updatedAttributes", "type": { "type": "map", "values": "string" } },
{ "name": "previousAttributes", "type": { "type": "map", "values": "string" } },
{ "name": "actorHostname", "type": ["null", "string"] },
{ "name": "contentURI", "type": ["null", "string"] },
{ "name": "previousContentURI", "type": ["null", "string"] },
{ "name": "parentIds", "type": { "type": "array", "items": "string" } },
{ "name": "childIds", "type": { "type": "array", "items": "string" } },
{ "name": "platform", "type": "string" },
{ "name": "application", "type": "string" },
{ "name": "remoteIdentifier", "type": ["null", "string"] },
{ "name": "alternateIdentifier", "type": ["null", "string"] },
{ "name": "transitUri", "type": ["null", "string"] }
]
}
</code>
</pre>
</body>
</html>