Timeout issue with RA3 nodes

[ryancharris]

Describe the bug

Filing on behalf of a user who contacted customer support:

As this timeout is happening while opening the connection, the team do not see much on the Redshift logs indicating that the issue could be from Client side where some setting is causing the connection to end prematurely. Can you please review the TCP keepAlive settings on the client that was connecting to these sessions and see if that mitigates the issue?

Steps To Reproduce

N/A

Expected behavior

There should not be a timeout with RA3 nodes.

Screenshots and log output

N/A

System information

Which database are you using dbt with?

• postgres
• redshift
• bigquery
• snowflake
• other (specify: ____________)

The output of dbt --version:

<output goes here>

The operating system you're using:

The output of python --version:

Additional context

Zendesk ticket

[jtcohen6]

Thanks for opening @ryancharris! I don't yet have a good sense of what's going on here. I'm hoping to dive into the logs in the linked ticket (once I get access) to get a better sense of what's going on.

The way dbt connects to RA3-node Redshift clusters is the same way it connects to all other Redshift clusters/node types, which is the same way dbt connects to Postgres. There is a keepalives_idle profile parameter that maps to the keepalives_idle Postgres parameter (docs), which is useful for long-running queries. By default, it's set to 0, i.e. the system default.

Removing the triage tag while waiting on more information.

[jtcohen6]

We never got to the bottom of this, and there isn't a lot to go on. I'm going to close in favor of #3303, since that seems like our best chance at a workable resolution for intermittent (and otherwise inexplicable) timeouts