Comparing "pg_controldata" output on prod and standby might help you with this.
We do use this approach and it is pretty reliable and gives time lag up to the granularity of checkpoint_timeout.
On Thu, Feb 23, 2012 at 11:51 AM, Stuart Bishop <stuart@xxxxxxxxxxxxxxxx> wrote:
On Thu, Feb 23, 2012 at 2:58 AM, Greg WilliamsonMight help indeed. My existing solution already has a small daemon (I
<gwilliamson39@xxxxxxxxx> wrote:
> Stuart Bishop shaped the aether to ask:
>
>> Hi.
>>
>> I need to measure how far in the past a hot standby is, async
>> streaming replication.
>
> Not sure if this will help, but we are using repmgr <https://github.com/greg2ndQuadrant/repmgr>; it sets up a monitoring schema which we poll )see the "Monitoring and Testing" section ... study their source code some and see how they come up with lag times.
can't always query the Slony-I sl_status view fast enough for load
balancing web requests, so I maintain a cache). But repmgr seems to
cover other work I need to do to keep ops happy so something for me to
look closer at.
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
