This is a tracking issue for writing roachtests that setup different failure scenarios to then motivate defining, implementing, and documenting how C2C behaves in the face of such failures:

• Single node failure in the primary cluster
• Single node failure in the standby cluster
• Temporary network outage (and/or latency) between the primary and standby
• Temporary load increase on the primary
• Significant change in the topology of the primary cluster
• Significant change in the topology of the standby cluster
• Range becomes unavailable on the primary cluster
• Node runs out of disk on the destination cluster

We expect each of these scenarios to result in the discovery of more work required to make C2C resilient, these will be tracked in separate issues.

Epic: CRDB-18753

Jira issue: CRDB-20281