Operations
Health checks, logs, backups, getting support, and removing an install.
Health and logs
- Health:
https://APP_HOST/api/healthanswers"ok": truewhen the web app is serving. The same answer reports the engine connection (containerOk) and the cache separately, so check those too. - Logs: the app and MCP endpoint log to the cluster (
kubectl -n eode logs deploy/eode-app). That needs AdminPrincipalArn set on the stack and a network path to the cluster's API, which is private by default. Install and upgrade logs are in CloudWatch Logs, log group/aws/codebuild/<cluster>-installer.
Backups
| What | How it's backed up |
|---|---|
| Database | RDS automated backups, kept DbBackupDays days, plus a snapshot before every stack update |
| Datasets, results, conversations | In the database, covered by the backups above |
| Starter library and engine working folders | The EFS file system, with automatic backups on |
| Secrets | Secrets Manager. Keep eode/data-connections safe: without it, saved data-connection credentials can't be read |
To restore, contact River Logic. The stack can't switch to a restored database by itself, so we restore the database and file system together with you.
Support
Contact your River Logic contact. Say which release you run and what happened. eode sends no diagnostics or logs to River Logic; anything we see, you send us.
Uninstall
- If you might restore later, copy the value of
eode/data-connectionsfrom Secrets Manager somewhere safe. The stack deletes its secrets, and saved data-connection credentials can't be read without that one. - Turn off deletion protection on the eode database in RDS.
- Delete the stack in CloudFormation.
The database leaves a final snapshot, the snapshots taken before each upgrade (eode-pre-…) stay, and the EFS file
system and its backups are kept. They're billed until you delete them in RDS, EFS and AWS Backup. Ask us to remove
your account's invitation when you're done.