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
Previously, tsdump upload to Datadog was taking more time compare to stage
tsdump in roachprod. This was inadequate because it would increase the MTTD
(Mean Time To Detect) an issue. This change introduces `upload-workers` as flag
to set the number of Datadog upload workers. The default value is 50. This
change includes the changes around the retry configuration to further improve
performance.
Epic: None
Part of: CRDB-52094
Release note: None
Copy file name to clipboardExpand all lines: pkg/cli/debug.go
+1Lines changed: 1 addition & 0 deletions
Original file line number
Diff line number
Diff line change
@@ -1617,6 +1617,7 @@ func init() {
1617
1617
f.StringVar(&debugTimeSeriesDumpOpts.userName, "user-name", "", "name of the user to perform datadog upload")
1618
1618
f.StringVar(&debugTimeSeriesDumpOpts.storeToNodeMapYAMLFile, "store-to-node-map-file", "", "yaml file path which contains the mapping of store ID to node ID for datadog upload.")
1619
1619
f.BoolVar(&debugTimeSeriesDumpOpts.dryRun, "dry-run", false, "run in dry-run mode without making any actual uploads")
1620
+
f.IntVar(&debugTimeSeriesDumpOpts.noOfUploadWorkers, "upload-workers", 50, "number of workers to upload the time series data in parallel")
0 commit comments