Failed to get catalog items

I’m getting this alert occasionally (about 5 times every 3 months). I find it using the AWS console search for alerts.

Failed to get catalog items when cleaning catalog cache

The ‘Cause’ is usually different, but one of the following:

1 - Read timed out
2 - Temporary failure in name resolution
3 - Internal server error (Service: AmazonDynamoDBv2; Status Code: 500; Error Code: InternalServerError; Request ID: A922QM5HJQB92POORRB849UJ1VVV4KQNSO5AEMVJF66Q9ASUAAJG)

Which leads to this error during the transaction:

{:cognitect.anomalies/category :cognitect.anomalies/unavailable, :cognitect.anomalies/message "Loading database", :context nil}


Hi @jarrod,

What version of Datomic Cloud are you using and what topology? Is your system creating or deleting DBs as part of any test case?


I’m running the solo topology.

ComputeCFTVersion 303
DatomicCloudVersion 8295


Did you happen to typo that version? I think the only 303 release we have is 303-8300, which was our second release ever. If the version is correct, in the sense that you are running a release from 2018 --you should upgrade as this issue is likely resolved on the latest. We have had a few issues in the interim releases which could be related.

I recommend upgrading to the latest.

That is correct @jaret , I believe it is the second release. Which is a kudos to the team, still running solidly. Although, I hear you - I do need to upgrade it. I’m a little worried to upgrade. We’ve been kicking the can and now I’m concerned that it may be difficult. Are there any reasons to believe that it will not be straightforward, given that it was one of the earlier versions?

@jarrod While I don’t suspect there will be major issues upgrading, we’d be happy to help guide the upgrade. If you want to open a support case shoot us an e-mail to We can gather the information we need there to help confirm the proper upgrade path.

I am concerned about your current system in regards to the critical releases:

I am also surprised that your system is running at all given that the version of Node.js used in the lambdas on the older versions of Datomic Cloud was discontinued by AWS. See