Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Migrated to Confluence 5.3

Apache Ranger by design provides configurable audits destinations. One of destinations is slf4j logging interface, which means audits can be streamed into any logging framework that is bound to slf4j. With this feature, Ranger can support streaming audits into popular messaging bus e.g. Kafka to provide realtime data feeds for various monitoring system.

Enabling audits logging is as simple as adding some configurations in logging properties files and adding a few kafka libraries.

Here is an example of configuring Hive logging properties file to enable sending audit events to Kafka messaging bus. 

Code Block
# kafka Appender
### for Ranger 0.4.0,KAFKA_HIVE_AUDIT
log4j.appender.KAFKA_HIVE_AUDIT.layout.ConversionPattern=%d{ISO8601} %-5p [%t]: %c{2} (%F:%M(%L)) - %m%n

For Kafka 0.8.1.x, the following jar files should be put into $HIVE_DIR/lib

Code Block
titlekafka libraries


In consumer side, some deserializer should be there to parse normalized audit records

Code Block
titlehive audit parser
public class AuditLogJsonDeserializer extends JsonDeserializer<HiveAuditLogDataModel> {
	private static Logger LOG = LoggerFactory.getLogger(AuditLogJsonDeserializer.class);
	public HiveAuditLogDataModel deserialize(JsonParser jp, DeserializationContext ctxt) 
			throws IOException, JsonProcessingException {
		HiveAuditLogDataModel model = new HiveAuditLogDataModel();
		JsonNode node = jp.getCodec().readTree(jp);
		String resource = node.get("resource").asText();
		// split resource to database, table, and column
		String[] tmp = resource.split("/");
		if(tmp.length >= 1){
			model.db = tmp[0];
			if(tmp.length >= 3){
				model.table = tmp[1];
				model.column = tmp[2];
		model.action = node.get("action").asText();
		model.clientIP = node.get("cliIP").asText();
		SimpleDateFormat formatter = new SimpleDateFormat("yyyyMMdd-HH:mm:ss.SSS-Z");
			model.timestamp = formatter.parse(node.get("evtTime").asText()).getTime();
		}catch(Exception ex){
			LOG.error("fail converting evtTime in hive audit log", ex);
		model.user = node.get("reqUser").asText();
		return model;