I would like long queries to kill themselves after ~ 30 seconds, in case of unexpected scenarios of growing load on the server.
It already happened that my whole cluster was under heavy load with a growing pile of search tasks (by checking GET _tasks
).
I am testing the timeout parameter like this
POST index_name/_search?timeout=1nanos
{
"size": 10000
}
but I get in the result:
{
"took": 38,
"timed_out": false,
}
I also tried with heavier aggregation that took 10 seconds and it doesn't time out.
- What does the timeout parameter do?
- Is there any way to accomplish this?