Feedback-ID: i8db94900:Fastmail
MIME-Version: 1.0
Date: Sat, 17 Jan 2026 21:52:54 +0100
From: "Lillian Berry" <lillian@star-ark.net>
To: "Andrei Lepikhov" <lepihov@gmail.com>,
 pgsql-performance@lists.postgresql.org
Cc: infra@pluralkit.me
Message-Id: <a985f55c-24fa-45aa-b373-8b30bc1e1fd3@app.fastmail.com>
In-Reply-To: <1d79d974-acfe-483c-81da-61989ed2f91b@gmail.com>
References: <aaf503fd-c08c-465d-9d80-4a71abb722c1@app.fastmail.com>
 <1d79d974-acfe-483c-81da-61989ed2f91b@gmail.com>
Subject: Re: Slow queries on simple index
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/a985f55c-24fa-45aa-b373-8b30bc1e1fd3%40app.fastmail.com>
Precedence: bulk

> Hmm, where is the evidence that your query uses theindex? Maybe the 
> generic plan accidentally forces SeqScan?

You may be correct - I just noticed that I am able to reproduce slow queries through my app. However, running the exact same queries in psql does use the index.

What might be different about my app's connections than psql? It uses Npgsql 9.0.2 on .NET 8, with these settings:

// src is "Host=[fdaa:9:e856:a7b:910f:0:a:102];Port=5432;Username=pluralkit;Database=pluralkit;Maximum Pool Size=150;Minimum Pool Size = 50;Max Auto Prepare=50"
var builder = new NpgsqlConnectionStringBuilder(src)
{
    Pooling = true,
    Enlist = false,
    NoResetOnClose = true,
    Timeout = 2
};

My understanding is slow query logs calculate the query time on the backend, and any client or network latency is not counted. Is this correct?

Kindly,
Lillian