We are using Glue Crawler and switched to the S3 event notification 2 month or so ago. It has been failing with Internal Service exception fairly regularly so we had to switch back to scan all folders, run the crawler that way and then switch back to S3 event notification again. Previously problem seemed to happen when there were many events in SQS (40k+). S3 event based would work ok on the smaller amounts.
Since the start of the 2022 year, event based crawler has been constantly failing with Internal Service Exception. Even after successfully running the all folder scan and switching back to the event based with just over 1k of events, it is still failing. No other errors provided in the log.
The S3 full folder scan Crawler runs successfully however it takes 6+ hours to complete which is not acceptable for us.
I did check https://aws.amazon.com/premiumsupport/knowledge-center/glue-crawler-internal-service-exception/ for this issue however none of the mentioned conditions exist in our case and as I mentioned the full scan crawler runs without issues.
Log:
INFO : Crawler configured with Configuration
{
"Version": 1,
"CrawlerOutput": {
"Partitions": {
"AddOrUpdateBehavior": "InheritFromTable"
}
},
"Grouping": {
"TableGroupingPolicy": "CombineCompatibleSchemas",
"TableLevelConfiguration": 2
}
}
and SchemaChangePolicy
{
"UpdateBehavior": "UPDATE_IN_DATABASE",
"DeleteBehavior": "DEPRECATE_IN_DATABASE"
}
ERROR : Internal Service Exception
Any help would be appreciated.