Crawl log - errors

Question

I read many articles but I didn't find a solution for my problem.

Our SP 2019 on-prem farm consists of: 2x Application with Search, 2x Front-end with Distributed Cache. Each server have 32GB RAM and 8 virtual processors.

Full crawl showed me the following errors:

When running a full crawl, the load is as follows:

Search Application Topology:

Error voices like "Processing this item failed" should be related to a high load CPU/RAM. RAM is still < 80%. I have set Performancelevel for search service: PartlyReduced. I increased the timeout value under Farm Search Administration to 600 from 60. Content source contains only one web application. Full Crawl takes more than 5 hours for this content source.

Questions:

1. How do I get rid of these "Processing ..." error voices? (The possibility would be to add another Application with Search server?)
2. "SharePoint returned an empty response" - There are files with a size of 0KB. Is it possible to add a crawl rule to exclude?
3. "The object was not found" - After loading the url at which this error message is displayed, I get: Page Not Found. Index Reset doesn't work on this problem.

I will be grateful if you help me solve these problems.

Answer 1

Ultimately the graph isn't lying, you need a higher core count (also using more than 1 socket serves no purpose in a SharePoint VM as SharePoint isn't licensed by socket). In addition, you shouldn't ever be scheduling full crawls. Full crawls are for initial crawls or to resolve an index that is otherwise unrecoverable. Use continuous crawls only.

OOM doesn't necessarily mean the system is out of memory itself, there are other factors such as memory that is able to be allocated to the process, but likely increasing the amount of memory will alleviate this, too.

I wouldn't be concerned with attempting to correct all crawl errors (such as #2 and #3).