Hello. Our Zendesk Support jobs take around 1 hour...
# replication-troubleshooting
m
Hello. Our Zendesk Support jobs take around 1 hour to complete even though I configured the streams as incremental and the size of our data is relatively small (around 3 million tickets). I wonder if anyone else has the same problem.
Our EC2 has 32 GiB of memory and 8 vCPU. Besides the other connectors run very fast.
m
Murat do you mind sharing the sync log or the sync summary with the records/time consumed? I didn’t find any issue in Github about Zendesk Support performance though
m
Hi @Marcos Marx (Airbyte). Here's the sync summary logs:
m
@Murat Cetink sorry to ask you but can you share the complete log file? Maybe the most time consuming operation is during normalization.
m
@Marcos Marx (Airbyte) we don't do any normalization, it's just raw load. Here's the log.
c
We are experiencing the same thing with the Zendesk --> Snowflake; running incremental takes an hour
m
From logs:
Copy code
2023-02-07 16:41:56 [44msource[0m > SourceZendeskSupport runtimes:
Syncing stream groups 0:00:00.926138
Syncing stream organizations 0:00:01.216432
Syncing stream ticket_audits 1:09:22.084443
Syncing stream ticket_comments 0:00:20.628914
Syncing stream ticket_fields 0:00:00.495242
Syncing stream tickets 0:00:11.601779
Syncing stream users 0:00:08.868781
the ticket_audit is consumit most of the time. Found this issue about `ticket_audits`https://github.com/airbytehq/airbyte/issues/16162 probably this endpoint can be improved… I’ll dig into the code after
👍 1
c
Ticket_audits 58 mins
u
I created a Github issue to investigate any problem with performance or improve our docs explinaing why this endpoint is so slow: https://github.com/airbytehq/airbyte/issues/22611
👍 2