# Cloud remediation with approvals built in | ZopNight

> 99 rules can be fixed from the product: 20 in one click, 79 after a guided review. ZopNight checks, gets approval, makes the change and validates it.

Source: https://zop.dev/zopnight/auto-remediation

---

_Check, approve, change, validate_

## Remediation, with approvals built in.

Click Remediate on a finding, or Review and remediate for a guided fix, and ZopNight runs the change for you: a check that it is still safe, an admin's approval where one is needed, the cloud action itself, then a validation that it worked. 99 rules can be fixed from the product: 20 in one click, 79 after a guided review.

- 99 — rules you can fix in the product
- 20 — fix in one click
- 79 — after a guided review
- Every — fix checked after it runs

Free to start. No card. The playground just needs your work email.

## A change you can stand behind.

Up to four steps, and the last one proves the others worked.

- 01. Check first — A precondition confirms the resource is still in the state the finding described before anything changes.
- 02. Get approval — When someone other than an admin starts a pause or a delete, an admin approves it first. Guided fixes add a type-to-confirm review on top.
- 03. Make the change — Pause, resize, delete, apply a lifecycle policy, or patch a live Kubernetes manifest, whichever the finding calls for.
- 04. Prove it took — A validation step confirms the change landed. If it failed, the error says whether it is a permission to grant, a retry to try, or something to report.

## Hard lines it will not cross.

Speed matters. So does not breaking production.

- Databases: RDS, Aurora, Cloud SQL, ElastiCache, Azure SQL, Postgres and MySQL are never resized, paused or deleted. The most a fix does there is change backup settings, after a guided review.
- No one-shot stops: A fix never stops a VM on its own. Savings from switching things off at night go to a schedule instead.
- Singletons: A Kubernetes workload you deliberately run as one replica is never scaled automatically. That fix is guided, and you confirm it.
- Audit trail: Every step lands in the audit log, which admins can export to CSV. Where no safe change exists, the finding carries a numbered how-to-fix playbook instead of a button.

## Stop filing tickets for fixes a button can do.

Connect an account with Read + Write access and the first fixable finding is a click away.

[object Object]