Retriable Error "KafkaJSRequestTimeoutError" was thrown uncaught during a rebalance

[ModernTrollfare]

Describe the bug
A "KafkaJSRequestTimeoutError" was thrown, uncaught in our user App. It was marked "Retriable: true", but seemingly no auto-retry was carried out.

To Reproduce
There are no clear steps to reproduce this, but the following is observed:

1. A rebalance was triggered. Our Topic has exactly 1 Consumer per group as each consumer was assigned a unique Consumer Group ID during KafkaJS instance creation, and they have exactly 1 partition. Thus, we assume that a Heartbeat Timeout is triggered.
2. KafkaJS reported error "This is not the correct coordinator for this group".
3. KafkaJS reported warn "The group is rebalanceing, re-joining".
4. After a few seconds, the error is thrown:
   KafkaJSRequestTimeoutError: Request SyncGroup(key: 14, version: 3) timed out.

The timeouts used in KafkaJS:

connectionTimeout: 5000
requestTimeout: 10000
maxRetryTime: 30000
retries: 3

restartOnFailure is default (returning true)

Expected behavior
The library retries connection and re-joins indefinitely without throwing an uncaught error.

Observed behavior
An uncaught error marked "retriable: true" was thrown, and crashes the application.
Logs:
2022-07-08T01:13:07.923Z ERROR [Connection] Response OffsetCommit(key: 8, version: 5) (meta:{"level":"error","logger":"kafkajs","broker":"<broker>:9092","clientId":"<clientID>","error":"This is not the correct coordinator for this group","correlationId":14179,"size":78})
2022-07-08T01:13:07.925Z WARN [Runner] The group is rebalancing, re-joining (meta: {"level":"warn","logger":"kafkajs","groupId":"<unique groupID>","memberId":"<memberID>","error":"This is not the correct coordinator for this group"})
Error thrown:
KafkaJSRequestTimeoutError: Request SyncGroup(key: 14, version:3) timed out at SocketRequest.timeoutRequest (node_modules/kafkajs/src/network/requestQueue/socketRequest.js:107:19) at node_modules/kafkajs/src/network/requestQueue/index.js:93:21 at Map.forEach (<anonymous>) at Timeout._onTimeout (node_modules/kafkajs/src/network/requestQueue/index.js:91:23) at listOnTimeout (node:internal/timers:568:17) at processTimers (node:internal/timers:510:7) { retriable: true, helpUrl: undefined, broker: <broker>, correlationId: 3, createdAt: <epoch timestamp>, sentAt: <epoch timestamp>, pendingDuration: 0 }

Environment:

• KafkaJS v2.0.1
• Ubuntu 21
• nodejs v18.2
• Kafka is hosted on Confluent cloud.
• One single broker specified in application.

[ModernTrollfare]

Seems to be resolved with the fix mentioned in #1382 and #1384.

The current observation on v2.1.0:

• Retrier correctly catches a throw during reconnect attempts and fails with a line of log instead of crashing the whole app.
• The 'crash' during retry is catchable through instrumentation event interface.