zeroclaw-labs/zeroclaw · error

Qdrant collection info failed ({status}): {text}

Error message

Qdrant collection info failed ({status}): {text}

What it means

count() GETs /collections/{c} and reads result.points_count; purge_session_for_agent() calls it after purging. A non-2xx on that GET raises this error.

Source

Thrown at crates/zeroclaw-memory/src/qdrant.rs:897

            .len())
    }

    async fn count(&self) -> Result<usize> {
        self.ensure_initialized().await?;

        let resp = self
            .request(
                reqwest::Method::GET,
                &format!("/collections/{}", self.collection),
            )
            .send()
            .await
            .context("failed to get Qdrant collection info")?;

        if !resp.status().is_success() {
            let status = resp.status();
            let text = resp.text().await.unwrap_or_default();
            anyhow::bail!("Qdrant collection info failed ({status}): {text}");
        }

        let json: serde_json::Value = resp.json().await?;

        let count = json
            .get("result")
            .and_then(|r| r.get("points_count"))
            .and_then(|c| c.as_u64())
            .unwrap_or(0);

        let count =
            usize::try_from(count).context("Qdrant returned a points count that exceeds usize")?;
        Ok(count)
    }

    async fn health_check(&self) -> bool {
        let resp = self.request(reqwest::Method::GET, "/").send().await;

View on GitHub (pinned to 88bb9c8533)

Solutions

1. Verify collection presence and key with curl
2. If dropped: recreate and restart the process (re-init)
3. 503: retry after optimization completes (check the collection status field)
4. Treat count as advisory at the call site so a failed count doesn't discard the purge result

fn is_qdrant_collection_info_error(e: &anyhow::Error) -> bool {
    e.to_string().starts_with("Qdrant collection info failed")
}

let n = match memory.count().await {
    Ok(n) => n,
    Err(e) if is_qdrant_collection_info_error(&e) => 0, // stats are advisory; keep the purge result
    Err(e) => return Err(e),
};

When it happens

Trigger: 404 collection deleted after init; 401/403 bad key; 503 while Qdrant is resharding or optimizing segments.

Common situations: A post-purge stats call hitting a collection that ops just dropped; api-key rotation; Qdrant returning 503 during optimizer runs.

Related errors

• Qdrant collection creation failed ({status}): {text}
• Qdrant delete failed ({status}): {text}
• Qdrant search failed ({status}): {text}
• Qdrant set payload failed during agent rename ({status}): {t
• Qdrant upsert failed ({status}): {text}

AI-assisted analysis of zeroclaw-labs/zeroclaw@88bb9c8533 (2026-08-23). Data as JSON: /api/errors/64d5963fd9548f7c. Report an issue: GitHub.