On Mon, Jan 5, 2015 at 9:56 PM, Peter Eisentraut <pete...@gmx.net> wrote:
> On 11/19/14 7:36 AM, Magnus Hagander wrote: > > + <row> > > + > > <entry><structname>pg_stat_ssl</><indexterm><primary>pg_stat_ssl</primary></indexterm></entry> > > + <entry>One row per connection (regular and replication), showing > statistics about > > + SSL used on this connection. > > + See <xref linkend="pg-stat-ssl-view"> for details. > > + </entry> > > + </row> > > + > > It doesn't really show "statistics". It shows information or data. > Good point. We should make contrib/sslinfo a wrapper around this view as much as > possible. > Agreed - but let's do that as a separate patch. Is it useful to include rows for sessions not using SSL? > I think so, mainly because it makes things "more obvious" that you are querying it the right way. Sure you could do a LEFT JOIN from pg_stat_activity and a CASE to get the same results, but that makes it a lot less in-your-face. Should we perpetuate the "ssl"-naming? Is there a more general term? > tls? :) We call it ssl everywhere else, I think being consistent is important. Will this work for non-OpenSSL implementations? > Yes, it uses (and declares new) the new internal APIs that Heikki created. -- Magnus Hagander Me: http://www.hagander.net/ Work: http://www.redpill-linpro.com/
