Share
Report
3We have recently been impacted by the deprecation of the Amazon Redshift connector in Microsoft Power Automate.
Context:
* Our Power Automate flows were previously using the native Amazon Redshift connector without issues.
* Since May 8th, all flows started failing on the very first GET action with HTTP 400 errors.
* We discovered that the Amazon Redshift connector is now deprecated, so we started migrating our flows to the PostgreSQL connector instead.
* Since May 8th, all flows started failing on the very first GET action with HTTP 400 errors.
* We discovered that the Amazon Redshift connector is now deprecated, so we started migrating our flows to the PostgreSQL connector instead.
Current situation:
* The PostgreSQL connection itself works successfully.
* We can authenticate and connect to the Redshift cluster.
* However, when configuring actions such as "Get rows", Power Automate cannot find our objects and returns:
* We can authenticate and connect to the Redshift cluster.
* However, when configuring actions such as "Get rows", Power Automate cannot find our objects and returns:
* HTTP 404
* "Object not found"
* "Object not found"
Important details:
* The objects absolutely exist in Redshift.
* They are mostly SQL views, not physical tables.
* The same objects were previously working through the Redshift connector.
* We strongly suspect these are Redshift views not properly exposed through PostgreSQL metadata discovery.
* They are mostly SQL views, not physical tables.
* The same objects were previously working through the Redshift connector.
* We strongly suspect these are Redshift views not properly exposed through PostgreSQL metadata discovery.
What we already tested:
* Using fully qualified names:
* schema.view_name
* Verifying permissions:
* Verifying permissions:
* GRANT USAGE ON SCHEMA
* GRANT SELECT ON VIEW/TABLE
* SSL enabled
* Same credentials/database/endpoint as before
* Confirmed the views are queryable directly in Redshift
* GRANT SELECT ON VIEW/TABLE
* SSL enabled
* Same credentials/database/endpoint as before
* Confirmed the views are queryable directly in Redshift
Main blocker:
Even if runtime execution might theoretically work, Power Automate marks the PostgreSQL action as invalid because of the 404 metadata lookup failure, and therefore the flow cannot be saved or validated.
Even if runtime execution might theoretically work, Power Automate marks the PostgreSQL action as invalid because of the 404 metadata lookup failure, and therefore the flow cannot be saved or validated.
Questions:
1. Has anyone successfully connected Power Automate PostgreSQL connector to Amazon Redshift views?
2. Are Redshift views officially unsupported by the PostgreSQL connector?
3. Is there any workaround to bypass metadata validation in the Power Automate designer?
4. Do we need to expose physical tables instead of views for compatibility?
5. Has anyone implemented a stable architecture pattern for Power Automate + Redshift after the connector deprecation?
2. Are Redshift views officially unsupported by the PostgreSQL connector?
3. Is there any workaround to bypass metadata validation in the Power Automate designer?
4. Do we need to expose physical tables instead of views for compatibility?
5. Has anyone implemented a stable architecture pattern for Power Automate + Redshift after the connector deprecation?
Our current fallback idea is:
* creating simplified physical tables dedicated to Power Automate consumption
* or introducing an API/Lambda layer between Power Automate and Redshift
* or introducing an API/Lambda layer between Power Automate and Redshift
Any feedback or experience would be greatly appreciated.
Categories:
I have the same question (0)
All responses (
Answers (
