nodefluent / kafka-streams

equivalent to kafka-streams :octopus: for nodejs :sparkles::turtle::rocket::sparkles:

Home Page:https://nodefluent.github.io/kafka-streams/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

prototype for sending successful messages to one topic, errors to another topic

dragonfriend0013 opened this issue · comments

I am trying to process a stream, and if the input processing errors out, i would like to send the data + error to a separate topic. good messages would be sent to another topic. Am i approaching this wrong?


import { KafkaStreams } from 'kafka-streams';

import getConfig from './config.mjs';
import getCache from './utils/redis.mjs';
import logger from './utils/logger.mjs';
import convertHrtime from './utils/convertHRTime.mjs';
import { tracer, getTraceId } from './utils/telemetry.mjs';
import formatData from './formatData.mjs';
import GetEvtQuery from './getEvents.mjs';

const config = getConfig();
const primaryTopic = process.env.KAFKA_INGRESS_TOPIC;
const retryTopic = process.env.KAFKA_RETRY_TOPIC;
const fmtTopic = `${primaryTopic}_FMT`;

const cache = getCache();
cache.on('ready', () => {
  logger.info('Redis is ready');
});
cache.on('error', (err) => {
  logger.error(`Redis error: ${err.code} ${err.syscall} ${err.address} ${err.port}`);
});

const kafkaStreams = new KafkaStreams(config);
kafkaStreams.on('error', (error) => {
  logger.error(`Kafka streams error: ${error.message}`);
});

const consume = async () => {
  const MFStream = kafkaStreams.getKStream();
  MFStream
    .from(primaryTopic)
    .from(retryTopic)
    .mapJSONConvenience()
    // eslint-disable-next-line consistent-return
    .map((message) => tracer.startActiveSpan('message', async (span) => {
      try {
        const start = process.hrtime();
        const msg = await formatData(message);
        msg.traceId = getTraceId(span);
        const evtConfig = await GetEvtQuery(msg.evtType, cache);
        if (!evtConfig) {
          return null;
        }
        // filter out configs that dont match filter
        // value will always be a string, so need to cast incoming data to string
        const events = evtConfig.filter((f) => f.keyFields
          .every((i) => msg.newMsg[i.field].toString() === i.value));
        msg.events = events;
        const duration = await convertHrtime(process.hrtime(start)).milliseconds;
        msg.durationFmtOpt = duration;
        span.setAttribute('duration', duration);
        span.end();
        return msg;
      } catch (e) {
         // check retryCnt, if >= 10, send to errortopic, otherwise send to retrytopic
const thisMsg = {
          ...message,
          errMsg: e.message,
        };
        let retryCnt = !thisMsg.retryCnt ? 0 : parseInt(thisMsg.retryCnt, 10);
        retryCnt += 1;
        thisMsg.retryCnt = retryCnt;
        if (thisMsg.retryCnt <= parseInt(process.env.MAX_RETRIES, 10)) {
          // await retryProducer(thisMsg);
        } else {
          // await errorProducer(thisMsg);
        }
        span.setStatus({
          code: SpanStatusCode.ERROR,
          message: e.message,
          willRetry: thisMsg.retryCnt <= parseInt(process.env.MAX_RETRIES, 10),
        });
        logger.error({
          msg: `${e.message.replace(/"/g, "'")}`,
          uuid: thisMsg.uuid,
          traceId: getTraceId(span),
          retryCnt: thisMsg.retryCnt,
        });
      }
    }))
    .awaitPromises()
    .wrapAsKafkaValue()
    .to(fmtTopic, 'auto', 'buffer');

  MFStream
    .start()
    .then(() => {
      logger.info('MFStream started.');
    }, (error) => {
      logger.error(`MFStream failed to start: ${error}`);
    });
};

consume().catch((e) => {
  // eslint-disable-next-line no-console
  console.error(e);
  logger.error(e);
  process.exit(1);
});

closing, used an example using getKafkaStyledMessage to override topic specified in stream.to