Examples with JsonByteArrayOperator org.apache.apex.malhar.lib.stream.JsonByteArrayOperator used on opensource projects

Example 1 with JsonByteArrayOperator

use of org.apache.apex.malhar.lib.stream.JsonByteArrayOperator in project apex-malhar by apache.

the class Application method populateDAG.

@Override
public void populateDAG(DAG dag, Configuration conf) {
    // set app name
    dag.setAttribute(DAG.APPLICATION_NAME, "Logstream Application");
    dag.setAttribute(DAG.STREAMING_WINDOW_SIZE_MILLIS, 1000);
    ConsoleOutputOperator mysqlConsole = dag.addOperator("MysqlConsole", ConsoleOutputOperator.class);
    ConsoleOutputOperator syslogConsole = dag.addOperator("SyslogConsole", ConsoleOutputOperator.class);
    ConsoleOutputOperator systemConsole = dag.addOperator("SystemlogConsole", ConsoleOutputOperator.class);
    LogScoreOperator logScoreOperator = dag.addOperator("logscore", new LogScoreOperator());
    /*
     * Read log file messages from a messaging system (Redis, RabbitMQ, etc)
     * Typically one message equates to a single line in a log file, but in
     * some cases may be multiple lines such as java stack trace, etc.
     */
    // Get logs from RabbitMQ
    RabbitMQLogsInputOperator apacheLogInput = dag.addOperator("ApacheLogInput", RabbitMQLogsInputOperator.class);
    RabbitMQLogsInputOperator mysqlLogInput = dag.addOperator("MysqlLogInput", RabbitMQLogsInputOperator.class);
    RabbitMQLogsInputOperator syslogLogInput = dag.addOperator("SyslogLogInput", RabbitMQLogsInputOperator.class);
    RabbitMQLogsInputOperator systemLogInput = dag.addOperator("SystemLogInput", RabbitMQLogsInputOperator.class);
    /*
     * Convert incoming JSON structures to flattened map objects
     */
    JsonByteArrayOperator apacheLogJsonToMap = dag.addOperator("ApacheLogJsonToMap", JsonByteArrayOperator.class);
    dag.addStream("apache_convert_type", apacheLogInput.outputPort, apacheLogJsonToMap.input);
    JsonByteArrayOperator mysqlLogJsonToMap = dag.addOperator("MysqlLogJsonToMap", JsonByteArrayOperator.class);
    dag.addStream("mysql_convert_type", mysqlLogInput.outputPort, mysqlLogJsonToMap.input);
    JsonByteArrayOperator syslogLogJsonToMap = dag.addOperator("SyslogLogJsonToMap", JsonByteArrayOperator.class);
    dag.addStream("syslog_convert_type", syslogLogInput.outputPort, syslogLogJsonToMap.input);
    JsonByteArrayOperator systemLogJsonToMap = dag.addOperator("SystemLogJsonToMap", JsonByteArrayOperator.class);
    dag.addStream("system_convert_type", systemLogInput.outputPort, systemLogJsonToMap.input);
    // operator for tuple counter
    Counter apacheLogCounter = dag.addOperator("ApacheLogCounter", new Counter());
    /*
     * operators for sum of bytes
     */
    SumItemFromMapOperator<String, Object> apacheLogBytesSum = dag.addOperator("ApacheLogBytesSum", new SumItemFromMapOperator<String, Object>());
    apacheLogBytesSum.setSumDimension("bytes");
    /*
     * opeartor filter 404 logs
     */
    SelectOperator filter404 = getFilteredMessagesOperator("Filter404", dag);
    /*
     * Explode dimensions based on log types ( apache, mysql, syslog, etc)
     */
    DimensionTimeBucketSumOperator apacheDimensionOperator = getApacheDimensionTimeBucketSumOperator("ApacheLogDimension", dag);
    DimensionTimeBucketSumOperator apacheFilteredDimensionOperator = getFilteredApacheDimensionTimeBucketSumOperator("ApacheFilteredLogDimension", dag);
    dag.addStream("apache_dimension_in", apacheLogJsonToMap.outputFlatMap, apacheDimensionOperator.in, apacheLogCounter.input, apacheLogBytesSum.mapInput, filter404.inport);
    dag.addStream("apache_filtered_dimension_in", filter404.outport, apacheFilteredDimensionOperator.in);
    DimensionTimeBucketSumOperator mysqlDimensionOperator = getMysqlDimensionTimeBucketSumOperator("MysqlLogDimension", dag);
    dag.addStream("mysql_dimension_in", mysqlLogJsonToMap.outputFlatMap, mysqlDimensionOperator.in);
    DimensionTimeBucketSumOperator syslogDimensionOperator = getSyslogDimensionTimeBucketSumOperator("syslogLogDimension", dag);
    dag.addStream("syslog_dimension_in", syslogLogJsonToMap.outputFlatMap, syslogDimensionOperator.in);
    // DimensionTimeBucketSumOperator systemDimensionOperator = getSystemDimensionTimeBucketSumOperator("systemLogDimension", dag);
    // dag.addStream("system_dimension_in", systemLogJsonToMap.outputMap, systemDimensionOperator.in);
    /*
     * Calculate average, min, max, etc from dimensions ( based on log types )
     */
    // aggregating over sliding window
    MultiWindowDimensionAggregation apacheMultiWindowAggCountOpr = getApacheAggregationCountOper("apache_sliding_window_count", dag);
    MultiWindowDimensionAggregation apacheMultiWindowAggSumOpr = getApacheAggregationSumOper("apache_sliding_window_sum", dag);
    dag.addStream("apache_dimension_out", apacheDimensionOperator.out, apacheMultiWindowAggCountOpr.data, apacheMultiWindowAggSumOpr.data);
    MultiWindowDimensionAggregation apacheFilteredMultiWindowAggCountOpr = getFilteredApacheAggregationCountOper("apache_filtered_sliding_window_count", dag);
    dag.addStream("apache_filtered_dimension_out", apacheFilteredDimensionOperator.out, apacheFilteredMultiWindowAggCountOpr.data);
    MultiWindowDimensionAggregation mysqlMultiWindowAggOpr = getMysqlAggregationOper("mysql_sliding_window", dag);
    dag.addStream("mysql_dimension_out", mysqlDimensionOperator.out, mysqlMultiWindowAggOpr.data);
    MultiWindowDimensionAggregation syslogMultiWindowAggOpr = getSyslogAggregationOper("syslog_sliding_window", dag);
    dag.addStream("syslog_dimension_out", syslogDimensionOperator.out, syslogMultiWindowAggOpr.data);
    // MultiWindowDimensionAggregation systemMultiWindowAggOpr = getSystemAggregationOper("system_sliding_window", dag);
    // dag.addStream("system_dimension_out", systemDimensionOperator.out, systemMultiWindowAggOpr.data);
    // adding top N operator
    TopN<String, DimensionObject<String>> apacheTopNCountOpr = dag.addOperator("apache_topN_count", new TopN<String, DimensionObject<String>>());
    apacheTopNCountOpr.setN(10);
    TopN<String, DimensionObject<String>> apacheFilteredTopNCountOpr = dag.addOperator("apache_filtered_topN_count", new TopN<String, DimensionObject<String>>());
    apacheFilteredTopNCountOpr.setN(10);
    TopN<String, DimensionObject<String>> apacheTopNSumOpr = dag.addOperator("apache_topN_sum", new TopN<String, DimensionObject<String>>());
    apacheTopNSumOpr.setN(10);
    TopN<String, DimensionObject<String>> mysqlTopNOpr = dag.addOperator("mysql_topN", new TopN<String, DimensionObject<String>>());
    mysqlTopNOpr.setN(5);
    TopN<String, DimensionObject<String>> syslogTopNOpr = dag.addOperator("syslog_topN", new TopN<String, DimensionObject<String>>());
    syslogTopNOpr.setN(5);
    // TopN<String, DimensionObject<String>> systemTopNOpr = dag.addOperator("system_topN", new TopN<String, DimensionObject<String>>());
    // systemTopNOpr.setN(5);
    /*
     * Analytics Engine
     */
    dag.addStream("ApacheLogScoreCount", apacheMultiWindowAggCountOpr.output, apacheTopNCountOpr.data, logScoreOperator.apacheLogs);
    dag.addStream("ApacheFilteredLogScoreCount", apacheFilteredMultiWindowAggCountOpr.output, apacheFilteredTopNCountOpr.data);
    dag.addStream("ApacheLogScoreSum", apacheMultiWindowAggSumOpr.output, apacheTopNSumOpr.data);
    dag.addStream("MysqlLogScore", mysqlMultiWindowAggOpr.output, mysqlTopNOpr.data, logScoreOperator.mysqlLogs);
    dag.addStream("SyslogLogScore", syslogMultiWindowAggOpr.output, syslogTopNOpr.data, logScoreOperator.syslogLogs);
    // dag.addStream("SystemLogScore", systemLogJsonToMap.outputMap, logScoreOperator.systemLogs);
    /*
     * Alerts
     */
    // TODO
    /*
     * Console output for debugging purposes
     */
    // ConsoleOutputOperator console = dag.addOperator("console", ConsoleOutputOperator.class);
    // dag.addStream("topn_output", apacheTopNOpr.top);entry.getValue()
    /*
     * write to redis for siteops
     */
    // prepare operators to convert output streams to redis output format
    AggregationsToRedisOperator<Integer, Integer> apacheLogCounterToRedis = dag.addOperator("ApacheLogCounterToRedis", new AggregationsToRedisOperator<Integer, Integer>());
    apacheLogCounterToRedis.setKeyIndex(11);
    AggregationsToRedisOperator<String, String> apacheLogBytesSumToRedis = dag.addOperator("ApacheLogBytesSumToRedis", new AggregationsToRedisOperator<String, String>());
    apacheLogBytesSumToRedis.setKeyIndex(9);
    AggregationsToRedisOperator<String, DimensionObject<String>> topNCountToRedis = getApacheTopNToRedisOperatorCountAggregation("topNCountToRedis", dag);
    AggregationsToRedisOperator<String, DimensionObject<String>> filteredTopNCountToRedis = getFilteredApacheTopNToRedisOperatorCountAggregation("filteredTopNCountToRedis", dag);
    AggregationsToRedisOperator<String, DimensionObject<String>> topNSumToRedis = getApacheTopNToRedisOperatorSumAggregation("topNSumToRedis", dag);
    // convert ouputs for redis output operator and send output to web socket output
    dag.addStream("apache_log_counter", apacheLogCounter.output, apacheLogCounterToRedis.valueInput);
    dag.addStream("apache_log_bytes_sum", apacheLogBytesSum.output, apacheLogBytesSumToRedis.valueInput);
    dag.addStream("topn_redis_count", apacheTopNCountOpr.top, topNCountToRedis.multiWindowDimensionInput, wsOutput(dag, "apacheTopAggrs"));
    dag.addStream("topn_redis_sum", apacheTopNSumOpr.top, topNSumToRedis.multiWindowDimensionInput, wsOutput(dag, "apacheTopSumAggrs"));
    dag.addStream("filtered_topn_redis_count", apacheFilteredTopNCountOpr.top, filteredTopNCountToRedis.multiWindowDimensionInput);
    // redis output operators
    RedisKeyValPairOutputOperator<String, String> redisOutTotalCount = dag.addOperator("RedisOutTotalCount", new RedisKeyValPairOutputOperator<String, String>());
    redisOutTotalCount.getStore().setDbIndex(15);
    RedisKeyValPairOutputOperator<String, String> redisOutTotalSumBytes = dag.addOperator("RedisOutTotalSumBytes", new RedisKeyValPairOutputOperator<String, String>());
    redisOutTotalSumBytes.getStore().setDbIndex(15);
    RedisMapOutputOperator<String, String> redisOutTopNCount = dag.addOperator("RedisOutTopNCount", new RedisMapOutputOperator<String, String>());
    redisOutTopNCount.getStore().setDbIndex(15);
    RedisMapOutputOperator<String, String> redisOutTopNSum = dag.addOperator("RedisOutTopNSum", new RedisMapOutputOperator<String, String>());
    redisOutTopNSum.getStore().setDbIndex(15);
    RedisMapOutputOperator<String, String> redisOutFilteredTopNSum = dag.addOperator("RedisOutFilteredTopNSum", new RedisMapOutputOperator<String, String>());
    redisOutFilteredTopNSum.getStore().setDbIndex(15);
    // redis output streams
    dag.addStream("apache_log_counter_to_redis", apacheLogCounterToRedis.keyValPairOutput, redisOutTotalCount.input);
    dag.addStream("apache_log_bytes_sum_to_redis", apacheLogBytesSumToRedis.keyValPairOutput, redisOutTotalSumBytes.input);
    dag.addStream("apache_log_dimension_counter_to_redis", topNCountToRedis.keyValueMapOutput, redisOutTopNCount.input);
    dag.addStream("apache_log_dimension_sum_to_redis", topNSumToRedis.keyValueMapOutput, redisOutTopNSum.input);
    dag.addStream("apache_log_filtered_dimension_count_to_redis", filteredTopNCountToRedis.keyValueMapOutput, redisOutFilteredTopNSum.input);
    /*
     * Websocket output to UI from calculated aggregations
     */
    dag.addStream("MysqlTopAggregations", mysqlTopNOpr.top, wsOutput(dag, "mysqlTopAggrs"), mysqlConsole.input);
    dag.addStream("SyslogTopAggregations", syslogTopNOpr.top, wsOutput(dag, "syslogTopAggrs"), syslogConsole.input);
    dag.addStream("SystemData", systemLogJsonToMap.outputMap, wsOutput(dag, "systemData"), logScoreOperator.systemLogs, systemConsole.input);
}

Example 2 with JsonByteArrayOperator

use of org.apache.apex.malhar.lib.stream.JsonByteArrayOperator in project apex-malhar by apache.

the class Application1 method populateDAG.

@Override
public void populateDAG(DAG dag, Configuration c) {
    int topNtupleCount = 10;
    LogstreamPropertyRegistry registry = new LogstreamPropertyRegistry();
    // set app name
    dag.setAttribute(DAG.APPLICATION_NAME, "Logstream Application");
    dag.setAttribute(DAG.STREAMING_WINDOW_SIZE_MILLIS, 500);
    RabbitMQLogsInputOperator logInput = dag.addOperator("LogInput", new RabbitMQLogsInputOperator());
    logInput.setRegistry(registry);
    logInput.addPropertiesFromString(new String[] { "localhost:5672", "logsExchange", "direct", "apache:mysql:syslog:system" });
    JsonByteArrayOperator jsonToMap = dag.addOperator("JsonToMap", new JsonByteArrayOperator());
    jsonToMap.setConcatenationCharacter('_');
    FilterOperator filterOperator = dag.addOperator("FilterOperator", new FilterOperator());
    filterOperator.setRegistry(registry);
    filterOperator.addFilterCondition(new String[] { "type=apache", "response", "response.equals(\"404\")" });
    filterOperator.addFilterCondition(new String[] { "type=apache", "agentinfo_name", "agentinfo_name.equals(\"Firefox\")" });
    filterOperator.addFilterCondition(new String[] { "type=apache", "default=true" });
    filterOperator.addFilterCondition(new String[] { "type=mysql", "default=true" });
    filterOperator.addFilterCondition(new String[] { "type=syslog", "default=true" });
    filterOperator.addFilterCondition(new String[] { "type=system", "default=true" });
    DimensionOperator dimensionOperator = dag.addOperator("DimensionOperator", new DimensionOperator());
    dimensionOperator.setRegistry(registry);
    String[] dimensionInputString1 = new String[] { "type=apache", "timebucket=s", "dimensions=request", "dimensions=clientip", "dimensions=clientip:request", "values=bytes.sum:bytes.avg" };
    // String[] dimensionInputString1 = new String[] {"type=apache", "timebucket=s", "dimensions=request", "dimensions=clientip","values=bytes.sum"};
    String[] dimensionInputString2 = new String[] { "type=system", "timebucket=s", "dimensions=disk", "values=writes.avg" };
    String[] dimensionInputString3 = new String[] { "type=syslog", "timebucket=s", "dimensions=program", "values=pid.count" };
    dimensionOperator.addPropertiesFromString(dimensionInputString1);
    dimensionOperator.addPropertiesFromString(dimensionInputString2);
    dimensionOperator.addPropertiesFromString(dimensionInputString3);
    LogstreamTopN topN = dag.addOperator("TopN", new LogstreamTopN());
    topN.setN(topNtupleCount);
    topN.setRegistry(registry);
    LogstreamWidgetOutputOperator widgetOut = dag.addOperator("WidgetOut", new LogstreamWidgetOutputOperator());
    widgetOut.logstreamTopNInput.setN(topNtupleCount);
    widgetOut.setRegistry(registry);
    ConsoleOutputOperator consoleOut = dag.addOperator("ConsoleOut", new ConsoleOutputOperator());
    dag.addStream("inputJSonToMap", logInput.outputPort, jsonToMap.input);
    dag.addStream("toFilterOper", jsonToMap.outputFlatMap, filterOperator.input);
    dag.addStream("toDimensionOper", filterOperator.outputMap, dimensionOperator.in);
    dag.addStream("toTopN", dimensionOperator.aggregationsOutput, topN.data);
    dag.addStream("toWS", topN.top, widgetOut.logstreamTopNInput, consoleOut.input);
    dag.setInputPortAttribute(jsonToMap.input, PortContext.PARTITION_PARALLEL, true);
    dag.setInputPortAttribute(filterOperator.input, PortContext.PARTITION_PARALLEL, true);
    dag.setInputPortAttribute(consoleOut.input, PortContext.PARTITION_PARALLEL, true);
}