You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment, benchmark generates two different flavors of an exported resources query when --catalog-query-pct is positive. Both of these filter on the resource type, without any consideration of the frequency of that type. In some testing with larger system (simulating 100k nodes), the Service type was frequently chosen for the query, a type that was extremely common, matching 11m rows in catalog_resources, causing an expensive full table scan, and a very large result.
Consider adjusting benchmark somehow, to prefer more selective exported resource queries.
The text was updated successfully, but these errors were encountered:
At the moment, benchmark generates two different flavors of an exported
resources
query when--catalog-query-pct
is positive. Both of these filter on the resourcetype
, without any consideration of the frequency of that type. In some testing with larger system (simulating 100k nodes), theService
type was frequently chosen for the query, a type that was extremely common, matching 11m rows incatalog_resources
, causing an expensive full table scan, and a very large result.Consider adjusting benchmark somehow, to prefer more selective exported resource queries.
The text was updated successfully, but these errors were encountered: