Error handling for publishing message to error exchange/queue

asserted the topic exchange for the publish channel as well

consumer/src/worker.js

@@ -37,6 +37,7 @@ export function initHandlers(handlers) {
  */
 export async function consume(channel, exchangeName, queue, publishChannel) {
   channel.assertExchange(exchangeName, 'topic', { durable: true });
+  publishChannel.assertExchange(exchangeName, 'topic', { durable: true });
   channel.assertQueue(queue, { durable: true });
   const bindings = _.keys(EVENT_HANDLERS);
   const bindingPromises = _.map(bindings, rk =>
@@ -78,11 +79,17 @@ export async function consume(channel, exchangeName, queue, publishChannel) {
           // we can use cloudamqp console to check the messages and may be manually create SF lead
           // nacking here was causing flood of messages to the worker and it keep on consuming high resources
           channel.ack(msg);
-          publishChannel.publish(
-            exchangeName,
-            EVENT.ROUTING_KEY.CONNECT_TO_SF_FAILED,
-            new Buffer(msg.content.toString())
-          );
+          try {
+            publishChannel.publish(
+              exchangeName,
+              EVENT.ROUTING_KEY.CONNECT_TO_SF_FAILED,
+              new Buffer(msg.content.toString())
+            );
+          } catch(e) {
+            // TODO decide if we want nack the original msg here
+            // for now just ignoring the error in requeue
+            logger.logFullError(e, `Error in publising Exchange to ${exchangeName}`);
+          }
         }
       }
     });

consumer/test/worker.spec.js

@@ -61,7 +61,8 @@ describe('worker', () => {
         },
       }, exchangeName, queueName,
       {
-        publish: channelPublishSpy
+        publish: channelPublishSpy,
+        assertExchange
       });
     }