On Thu, Jan 10, 2019 at 8:42 PM, Robert Hass wrote:
Thanks for comments.
>or at least not nearly as useful the results of a sampling approach.
I agree with your opinion.
Because it can't be asserted that the wait event is a bottleneck just because the number of wait event is large.
The same thing is mentioned in Oracle.
It also suggests that it is important to acquire waiting time for that.
----
https://docs.oracle.com/en/database/oracle/oracle-database/18/tgdba/measuring-database-performance.html#GUID-811E9E65-C64A-4028-A90E-102BBFF6E68F
5.2.3 Using Wait Events without Timed Statistics
----
>The data that LWLOCK_STATS produce are downright misleading
Is that so？
I do not know the need for this function.
---
Naoki Yotsunaga