NASA-IMPACT / csdap-cumulus

SmallSat Cumulus Deployment
Other
1 stars 1 forks source link

Upgrade Cumulus to 13.4.0 #85

Closed chuckwondo closed 1 year ago

chuckwondo commented 1 year ago

In Cumulus 13.4.0, it is now possible to configure memory sizes for lambda functions. Also, it is possible to set timeouts for lambda functions that were previously not exposed for such configuration.

We have a workaround that uses a Terraspace post-terraform-apply function for updating the configuration for some lambda functions to increase memory and timeout settings that could not be configured via terraform files prior to Cumulus 13.4.0 that should be removed as part of this upgrade.

kaulfusa commented 1 year ago

We have been encouraged not to upgrade Cumulus versions if there is no significant improvement to operational capability. 13.3 it appears will also be the minimum supported version of Cumulus for the foreseeable future as the Cumulus team is moving away from this concept.

Since we have a solution in place, I would suggest not upgrading at this time, or postponing the upgrade to later in the PI to prioritize Maxar data ingest. However, if this can be done prior to the next set of Maxar data being ready, then that is acceptable as well.

chuckwondo commented 1 year ago

@kaulfusa, thanks for the info on the new Cumulus upgrade policy.

I'd like to upgrade just to 13.4.0 (from 13.3.0), which should be extremely easy, and will allow us to remove some deployment "band-aids" because it fixes the issues that the "bands-aids" are for. This will eliminate some technical debt and slightly simplify deployment and eliminate a potential deployment failure point.

chuckwondo commented 1 year ago

@krisstanton, @jsrikish, and I have held 2 knowledge transfer sessions in the context of performing this upgrade. We plan to hold a 3rd session, which should allow us to wrap up this task.

chuckwondo commented 1 year ago

@jsrikish, @krisstanton, and I teamed up and finished upgrading my dev deployment and ran a smoke test confirming proper operation. I will wrap up work for creating a PR for this work so we can get the changes reviewed, merged, and deployed to our current UAT and Prod deployments.

chuckwondo commented 1 year ago

Successfully upgraded in dev, but reopening because deployments to UAT and Prod have not been completed yet.

chuckwondo commented 1 year ago

UAT deployment succeeded: https://github.com/NASA-IMPACT/csdap-cumulus/actions/runs/4049935613

chuckwondo commented 1 year ago

Prod upgrade is complete.