-t timeout in seconds (optional: default 10)
(If you specifies the maximum number of seconds to fulfill the request)
If you have a big range the value will have to be large.
In the UI, we break request for lots of data into smaller chunks. (85000 data points should be broken done by time, or as you found out will require a longer timeout)
1 of 1 people found this helpful
-i interval seconds (optional: default 0)
This is used when you specify the begin and end time (-b -e)
If this value is 0, a chart is assumed.
A chart has to be id fields.
If you specify -i interval with -b and -e, the result will be a series of drill down starting at begin time and progressing until end time.
it is good practice to break this up into reasonable begin and end time.
Thanks ... I guessed that -t was timeout but the results were strange
with a -t 30 I get 85940 row but with a -t 5 I get 17518 and a -t 10 I get 37265.. this seemed strange as no timeout messages were seen.
If I do not specify any -t I get 33579 which is around the 10 second timeout.
As I'm extracting in batch I think I'll set -t 60
Thanks for the prompt response (again)
The error messages would be on the agent.
The timeout is a select on a socket waiting for data on the GWS.
You may want to consider breaking the request into 12 hour or 8 hour chunks.
I believe from our UI testing that was the sweet spot for the best response.
If 60 second wait is acceptable (since it is batch), that works as well.