You’re hitting a known Redshift issue with Elementary’s

on-run-end

artifacts upload: Redshift (or the

redshift-connector

) rejects a single prepared statement that contains multiple commands (

begin; ... insert ...; commit;

), which is how the current

delete_and_insert

logic is compiled for Redshift. [[Slack thread](https://elementary-community.slack.com/archives/C08HW57FGN9/p1764339093392769); [GitHub issue](https://github.com/elementary-data/elementary/issues/2063)] ### Short‑term workarounds 1. Disable automatic artifacts upload in

on-run-end

In `dbt_project.yml`:

yaml    vars:      elementary:        disable_dbt_artifacts_autoupload: true        disable_run_results: true        disable_dbt_invocation_autoupload: true

This stops the failing

dbt_models

/

dbt_columns

uploads during

on-run-end

. You can then follow the docs to upload artifacts in a separate run instead of via hooks. [[GitHub issue](https://github.com/elementary-data/elementary/issues/2063)] 2. Temporarily disable Elementary entirely on problematic targets Example:

yaml    models:      elementary:        +schema: elementary        enabled: "{{ target.name in ['beta'] }}"

This was used by others to keep prod stable while debugging. [[GitHub issue](https://github.com/elementary-data/elementary/issues/2063)] 3. Macro override workaround (project‑level macro) One community member fixed it by overriding the Redshift implementation of

get_delete_and_insert_queries

so that it returns separate statements instead of a single multi‑command statement. [[Slack thread](https://elementary-community.slack.com/archives/C08HW57FGN9/p1764339093392769)] Create

macros/redshift_get_delete_and_insert_queries.sql

in your dbt project:

sql    {% macro redshift__get_delete_and_insert_queries(relation, insert_relation, delete_relation, delete_column_key) %}        {% set queries = [] %}         {% if delete_relation %}            {% set delete_query %}                delete from {{ relation }}                where                {{ delete_column_key }} is null                or {{ delete_column_key }} in (select {{ delete_column_key }} from {{ delete_relation }});            {% endset %}            {% do queries.append(delete_query) %}        {% endif %}         {% if insert_relation %}            {% set insert_query %}                insert into {{ relation }} select * from {{ insert_relation }};            {% endset %}            {% do queries.append(insert_query) %}        {% endif %}         {% do return(queries) %}    {% endmacro %}

This mirrors the community workaround and avoids

begin/commit

wrapping in a single prepared statement. [[Slack thread](https://elementary-community.slack.com/archives/C08HW57FGN9/p1764339093392769)] ### Status of an official fix The Elementary team acknowledged this Redshift issue and implemented a fix in the dbt package (PR

dbt-data-reliability#899

), with a note that it would be released “in a couple of days.” [[Slack thread](https://elementary-community.slack.com/archives/C08HW57FGN9/p1764339093392769); [GitHub issue](https://github.com/elementary-data/elementary/issues/2063)] However, the knowledge sources do not state clearly which released package version contains that fix, nor whether it’s backported to older versions. If you share your exact

elementary

dbt package version, you can: - Try upgrading to the latest available Elementary dbt package, or - Use the macro override +

disable_*_autoupload

vars above as a safe interim solution. If you’d like, paste your

packages.yml

and

vars: elementary:

section and I can suggest the minimal config change for your setup.

New message text here