Sometimes the clickhouse proxy server tries to redeploy itself after seemingly unrelated changes to other Terraform files. This is bad because the clickhouse proxy is:
- Hard to redeploy: redeploying the clickhouse proxy server requires you to manually run the clickhouse-proxy role in Ansible afterwards
- Very High Priority: Although the clickhouse proxy is not deployed in production right now, we are working to port the checking endpoint to production, which requires the clickhouse proxy. If it dies, we lose the check-in endpoint breaking all probes.
Given how important it is, we would like the clickhouse proxy to be redeployed only manually. A possible solution is to add a parameter to the ec2 module we use as base for the clickhouse proxy to specify whether a machine should be redeployed manually or automatically.
This is particularly important now that we need data from the check-in endpoint for the anonymous credentials project
Sometimes the clickhouse proxy server tries to redeploy itself after seemingly unrelated changes to other Terraform files. This is bad because the clickhouse proxy is:
Given how important it is, we would like the clickhouse proxy to be redeployed only manually. A possible solution is to add a parameter to the
ec2module we use as base for the clickhouse proxy to specify whether a machine should be redeployed manually or automatically.This is particularly important now that we need data from the check-in endpoint for the anonymous credentials project