ADR-0005: Be Resilient by Default
- Status: accepted
- Date: 2024-05-11
- Tags: quality, stability
Context and Problem Statement
When an VMware Secrets Manager component crashes or when an VMware Secrets Manager
component is evicted, the workloads can still function with the existing secrets
they have without having to rely on the existence of an active secrets store.
When an VMware Secrets Manager component restarts, it seamlessly recovers its
state from an encrypted backup without requiring manual intervention.
Decision Drivers
- Resilience is also related to being practically secure
- A resilient system is easy to operate, maintain, and troubleshoot.
- To have a highly-available system, we need to be resilient first.
Considered Options
- Be resilient by default.
- Think about resilience only when we have time.
Decision Outcome
Chosen option: “option 1”, because we cannot afford to have a system that
is not resilient.
Positive Consequences
Negative Consequences
- Additional work upfront.
- Additional complexity in the system.
ADRs
You can view the ADRs by browsing this following list:
edit this page ✏️