Okta Workflows: Cell Migration Expected Behaviors and Workarounds

The Okta Workflows team is executing a mandatory infrastructure migration from Redis versions 4 and 5 to Valkey for all clusters on the FL1 cell (including FL1-sch + DR, FL1-eng-rs, and FL1-loc). This change is required because AWS has ended standard support for Redis 4 and 5.

FL1 corresponds to Okta cells OK1, OK2, OK3, OK4, OK6, OK7, and OK11. See Determine the Okta Tenant or Cell Location for information on finding your Okta cell.

• Okta Workflows
• Customers on the FL1 cell

The migration is scheduled for Wednesday, July 22, 2026, starting at 9:00 AM UTC, and is expected to take between 3.5 and 4 hours, during which time customers will experience a temporary outage and failed workflow executions.

• No proactive configuration changes or immediate actions are required by customers prior to this migration window.
• During the 4-hour migration period, customers will experience a temporary service outage, including difficulties logging into Okta Workflows and receiving "Too many Cluster redirections" or timeout errors.
• Active workflow executions running during the maintenance window may fail or appear stuck "In Progress".
• If a customer's workflows fail to automatically refresh OAuth credentials post-migration, the admin must manually reauthorize the broken connections within the Okta Workflows Console.

What are the actual behaviors and expected errors that the customers will experience when the migration job is started?

During a previous migration of the Redis cluster to Valkey in the FL14 cell, there was a temporary outage. We anticipate there may be similar outages for customers on FL1.

FL1 corresponds to Okta cells OK1, OK2, OK3, OK4, OK6, OK7, and OK11. See Determine the Okta Tenant or Cell Location for information on finding your Okta cell.

The customer impact and associated behaviors included:

1. Workflow Access Issues: Customers experienced difficulties logging into Okta Workflows
2. Error Messages: Customers encountered errors such as "Too many Cluster redirections" and "there was an error changing the org". They may have also seen timeout errors or failed request errors
3. Failed Workflow Executions: Some workflow executions were affected and failed during the migration window. Please follow your normal process for any workflows that fail.
4. Credential Refresh Failures: Customers may have noticed that their flows did not automatically refresh OAuth credentials, requiring them to manually refresh expired credentials as a workaround

What should customers look out for after the migration?

1. Login/Access Issues: Temporary difficulties logging into Okta Workflows or encountering "There was an error changing the org" or timeout errors 
2. Throttled or Stuck Flows: Flows that appear stuck "In Progress" or fail to trigger because OAuth tokens failed to refresh automatically

What are the steps for the workaround if customers encounter errors?

If a customer encounters errors during or immediately after the migration, the primary workaround is to manually reauthorize/refresh their expired connection credentials within Okta Workflows.

Here are the step-by-step instructions for customers to resolve these issues.

1. Access the Console: Log in to the Okta Workflows Console. (If experiencing a temporary "Too many redirects" or login error, clear browser cache/cookies or try logging in again after a few minutes once the cluster stabilises).
2. Navigate to Connections: In the Workflows Console, go to the Connections tab.
3. Identify Broken Connections: Look for any connector or connection (such as the Okta connector or other 3rd-party OAuth connectors) showing an error state, or where flows are failing to execute due to expired credentials.
4. Reauthorize:
   1. Click on the affected connection.
   2. Select the Reauthorize (or Edit Connection) option.
   3. Enter the required credentials manually to re-establish the OAuth/API handshake.
5. Verify Flow Execution: Run a test to verify the impacted flows resume and execute normally.