[midPoint] AWS - RDS Postgres - DB - Very Slow
Pavol Mederly
mederly at evolveum.com
Thu Feb 4 10:55:37 CET 2021
Hi Gus,
> In the production environment with AWS RDS DB the timing of some
> actions seems to me to be very high.
Are you reading the right numbers? Because the most important are the
*avg:* values.
You can also (temporarily) enable recording values not only per
operation, but also per object type. So you will be able to compare e.g.
modifyObject times for e.g. user and task objects separately. Please see
*internals/repository/statistics/classification* sysconfig property and
its *perOperationAndObjectType* value.
Best regards,
Pavol
On 03/02/2021 14:35, Gus Lou via midPoint wrote:
> Hi Guys
> Some Information:
> Comparative - Midpoint Test vs Midpoint Prod
> From Midpoint -> Internals Configuration -> Performance
>
> Midpoint-Test (Midpoint App (01 node) and DB all single instance)
> Repository performance information:
> audit: count: 1, total time: 27 ms [min:
> 27, max: 27, avg: 27.0]
> cleanupAuditMaxAge: count: 1, total time: 386 ms [min:
> 386, max: 386, avg: 386.0]
> countContainers: count: 12, total time: 194 ms [min:
> 2, max: 97, avg: 16.2]
> countObjects: count: 6, total time: 281 ms [min:
> 4, max: 147, avg: 46.8]
> deleteObject: count: 1, total time: 303 ms [min:
> 303, max: 303, avg: 303.0]
> fetchExtItems: count: 1, total time: 62 ms [min:
> 62, max: 62, avg: 62.0]
> getObject: count: 390, total time: 12245 ms [min:
> 1, max: 5602, avg: 31.4]
> getVersion: count: 32, total time: 173 ms [min:
> 0, max: 23, avg: 5.4]
> modifyObject: count: 80, total time: 8911 ms [min:
> 7, max: 521, avg: 111.4], wasted time for 1 retry/retries: 276
> ms [min: 0, max: 276, avg: 3.5] ( 3.10%)
> searchContainers: count: 1, total time: 142 ms [min:
> 142, max: 142, avg: 142.0]
> searchObjects: count: 54, total time: 1737 ms [min:
> 1, max: 332, avg: 32.2]
> searchShadowOwner: count: 16, total time: 86 ms [min:
> 1, max: 13, avg: 5.4]
>
>
>
> Midpoint Prod (Midpoint App (02 Nodes) and AWD RDS DB)
> Repository performance information:
> countContainers: count: 20, total time: 80 ms [min: 1,
> max: 25, avg: 4.0]
> countObjects: count: 7, total time: 118 ms [min: 2,
> max: 97, avg: 16.9]
> getObject: count:1081559, total time: 2099910 ms [min: 1,
> max: 7652, avg: 1.9]
> getVersion: count: 51009, total time: 73060 ms [min: 0,
> max: 35, avg: 1.4]
> modifyObject: count: 47729, total time: 322292 ms [min: 5,
> max: 662, avg: 6.8]
> searchObjects: count: 52523, total time: 75512 ms [min: 1,
> max: 148, avg: 1.4]
>
> In the production environment with AWS RDS DB the timing of some
> actions seems to me to be very high.
> I continue to analyze, when I have news I share here.
> Regards
> Gus
>
>
> Em qua., 3 de fev. de 2021 às 09:22, Gus Lou <gugalou38 at gmail.com
> <mailto:gugalou38 at gmail.com>> escreveu:
>
> Hi Guys
>
> Does anyone here use Amazon RDS for PostgreSQL for midpoint database?
> I started testing this service and all midpoint access were very
> slow. I am investigating the cause, but if anyone has any
> experience in using database solutions as a service in conjunction
> with midpoint and can share I appreciate it.
>
> My Settings:
> Midpoint: 4.2
> Midpoint Nodes: 02
> OpenJDK: 11
> Postgres (AWS RDS): 11.8
> AWS Region: east2
> Instance Class: DB.m5.large
>
> Regards
>
> Gus
>
>
> _______________________________________________
> midPoint mailing list
> midPoint at lists.evolveum.com
> https://lists.evolveum.com/mailman/listinfo/midpoint
--
Pavol Mederly
Software developer
evolveum.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.evolveum.com/pipermail/midpoint/attachments/20210204/f7d2bbdd/attachment-0001.htm>
More information about the midPoint
mailing list