# What is the backup/restore best practice for integration with kots when using an existing cluster? **URL:** \ **Category:** Packaging an application\ **Tags:** kots\ **Created:** [October 10, 2022, 4:41pm UTC](https://community.replicated.com/t/what-is-the-backup-restore-best-practice-for-integration-with-kots-when-using-an-existing-cluster/1006 "2022-10-10T16:41:52Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![tomk](https://avatars.discourse-cdn.com/v4/letter/t/ee59a6/32.png) [@tomk](https://community.replicated.com/u/tomk)\ **Post date:** [October 10, 2022, 4:41pm UTC](https://community.replicated.com/t/what-is-the-backup-restore-best-practice-for-integration-with-kots-when-using-an-existing-cluster/1006/1 "2022-10-10T16:41:52Z")
Hi, I have a follow-up question about backup/restore after watching the following 2022 RepliCon Q3 talk: ``` https://www.youtube.com/watch?v=bpknVc1wrD8 minute 13:41 ``` We are preparing to deploy our first application in a variety of customer environments. Some may have the embedded cluster and some may not. Is there a good best-practice recipe that can be followed to achieve the nicely integrated Velero and kots support for all environments? (For example, it would seem we should include any third-party components like Velero itself as part of the application bundle and not as part of the kURL spec, since some environments won’t use kURL.) Thanks, Tom ---
**Author:** ![jdewinne](https://yyz2.discourse-cdn.com/flex030/user_avatar/community.replicated.com/jdewinne/32/152_2.png) [@jdewinne](https://community.replicated.com/u/jdewinne)\ **Post date:** [October 10, 2022, 7:53pm UTC](https://community.replicated.com/t/what-is-the-backup-restore-best-practice-for-integration-with-kots-when-using-an-existing-cluster/1006/2 "2022-10-10T19:53:48Z")
@tomk In the case of an existing k8s, the customer will have to install Velero. They can follow the [docs](https://docs.replicated.com/enterprise/snapshots-velero-cli-installing#install-the-velero-cli-in-an-online-cluster) to do so.