On 10/26/22 08:26, Yi Sun wrote:
>
>
> On Wed, 26 Oct 2022 at 18:10, jian he <jian(dot)universality(at)gmail(dot)com> wrote:
>
>
>
> On Wed, Oct 26, 2022 at 11:07 AM Yi Sun <yinan81(at)gmail(dot)com> wrote:
>
> Hi Guys,
>
> Who can help me with this please? I researched but still no
> result yet, thank you
>
> On Tue, 25 Oct 2022 at 16:30, Yi Sun <yinan81(at)gmail(dot)com> wrote:
>
> Hi,
>
> There are many databases in our production patroni cluster
> and it seems it is overloaded, so we decide to migrate the
> busiest database to a new patroni cluster.
>
> pgwatch2 is implemented, how to know how much CPU, RAM is
> used by the database please? Then we can use it to prepare
> the new patroni cluster hardware. Thank you
>
> Best regards
> Dennis
>
>
> manual:
> https://www.postgresql.org/docs/current/runtime-config-resource.html|
> |
> |except |min_dynamic_shared_memory| (|integer|)|
> |all other parameters are used to cap the memory. almost all
> parameters mentioned "database server" which means it's on cluster
> level.
> |
> |
> |
> |
> |
> --
> I recommend David Deutsch's <<The Beginning of Infinity>>
>
> Jian
>
>
>
> Hi Jian he
>
> Thank you for your reply
>
> The parameters are on cluster level, so we still can not know how much
> memory is used in a specific database, for example, total memory is 64GB
> 1. How to get how much memory is used on cluster level? For example 40GB
> 2. How to get how much memory is used in a specific database? For
> example 30GB, then we can prepare the new patroni cluster 32GB is enough
>
> Thank you
> Dennis
You can see connection with pg*backend* functions. You can log
connections to see which db is most commonly accessed. You can log sql
to see which table are being touched. You'll have to assume a
correlation to CPU/disc usage. What have you tried?