Skip to content
This repository has been archived by the owner on Jun 25, 2024. It is now read-only.

Setup operational concept with SRE #68

Closed
tobiscr opened this issue Aug 3, 2021 · 2 comments
Closed

Setup operational concept with SRE #68

tobiscr opened this issue Aug 3, 2021 · 2 comments
Assignees
Labels
area/lifecycle-management Used for any issues related to Kyma component lifecycle management area/reconciler required-for-kyma2.0

Comments

@tobiscr
Copy link
Contributor

tobiscr commented Aug 3, 2021

Description

The reconcile requires an operational concept which is aligned with SREs. The concept has to cover:

  • How the mothership- and component-reconciler(s) will be integrated into existing deployment pipelines (e.g. adding it pipelines which are also used to deploy KEB, provisioner etc.)
  • What the operational requirements for mothership- and component-reconcilers are (expected deliverables like trouble-shooting guide, further documentation, integration into SRE's logging-/monitoring-system)
  • What are the requirements / expectations from SRE side to the reconciler to analyse incidents (e.g. access to reconciler-logs, mandatory CLI features etc.): see Administrate reconciler remotely via CLI using REST API #115

AC:

  • Written concept (could also be a checklist or similar stored inside of this Github issue) which covers the agreed action items / decision and was reviewed by SREs
  • If tickets are created for some of these items, the tickets have to be referenced in the concept

Reasons

Ensure the reconciler is addressing operational constraints from SREs properly.

Attachments

@tobiscr tobiscr added the area/lifecycle-management Used for any issues related to Kyma component lifecycle management label Aug 3, 2021
@varbanv varbanv assigned varbanv and unassigned suleymanakbas91 Aug 26, 2021
@varbanv
Copy link
Contributor

varbanv commented Sep 7, 2021

After having a discussion with SREs on 02.09.2021 we came up with the following requirements:

  • Reconciler should be deployed using the same approach and as part of the KCP.
  • Reconciler database will be deployed and configured by SRE.
  • Reconciler will init and upgrade database using admin DB creadentials provided by SRE.
  • We will provide high level design documentation.
  • Metrics Exported will be used by SRE for monitoring.
  • Mothership and component reconciler logs will be scraped using Fluent Bit so CLI will not provide them.
  • KCP CLI will be augmented to include reconciler management functions: Administrate reconciler remotely via CLI using REST API #115
  • Reconciler will sync state with KEB to avoid data loss during down times: Reconciler has to synchronise cluster inventory with KEB #127

@tobiscr
Copy link
Contributor Author

tobiscr commented Nov 25, 2021

Agreed with @ebensom that all parts from reconciler side are considered

@tobiscr tobiscr closed this as completed Nov 25, 2021
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
area/lifecycle-management Used for any issues related to Kyma component lifecycle management area/reconciler required-for-kyma2.0
Projects
None yet
Development

No branches or pull requests

5 participants