Value cannot be null Eventhubs error

[NakulanR]

Hi Support,

We are seeing an issue in which data is not being replicated to Eventhubs. The Replicate logs show a series of errors like so: "[TARGET_APPLY ]I: Failed to produce Eventhubs message with record id <xx> to partition <-1> in hub '{hub-name}'. One or more errors occurred. Value cannot be null." The errors appear constantly as the task runs, however they don't cause the task to stop.

I'd like to understand 3 things:

• What is the "value" referencing in the error message? Is it a value in the data record or something in the connection that is Null?
• Is this data validation coming from Event Hubs or is Replicate able to identify an issue prior to sending the data to Event Hubs? i.e. Does Event Hubs reject the message or is Replicate blocking the message from being sent?
• Considering that data isn't being sent to Event Hubs, why is the error classed as an Info (I:) rather than an error message (E:)?

Any assistance with the above questions is greatly appreciated.

Thanks,

Nak

[DesmondWOO]

Hi @NakulanR ,

Thank you for reaching out to the Qlik Community.

Could you reproduce this issue with verbose logging enabled on TARGET_LOAD/TARGET_APPLY and check if any clues can be found in the task log? If you're not sure what's going wrong, please submit a support ticket.

Thanks,
Desmond

[NakulanR]

Hi @DesmondWOO,

Thanks for the prompt response. The trace level logging reports the following error: "inner exception is The request was terminated because the entity is being throttled. Error code : 50002. Sub error : 102. Please wait 4 seconds and try again". From my research it looks to be an issue with Event Hubs throttling, and increasing the throughput units (TU) on the hub instance may be a way to resolve the issue.

I will look into the issue on the Eventhubs side and reply with my findings

Regards,
Nak

[DesmondWOO · Accepted solution]

Hi @NakulanR ,

I reviewed other support cases, and it seems to be a capacity issue. Here are their suggested solutions:

• The customer successfully performed a full load for one table at a time, and the task worked perfectly.
• Temporarily increase the throughput units at the target endpoint while performing the full reload of the table.

Regards,
Desmond