docs/releases.md: Add info on cluster termination

equinor · Oct 9, 2018 · f375358 · f375358
1 parent 9269aa2
commit f375358
Showing 1 changed file with 11 additions and 8 deletions.
diff --git a/docs/releases.md b/docs/releases.md
@@ -4,15 +4,18 @@ This is an overview of current and previous Radix releases and deployments.
 
 We are experimenting with an approach with "release channels" which differ in update frequency, service level objectives(SLO) and agreements(SLA) and expected stability.
 
-|    Channel     |                                           Purpose                                                  |    Support    |      SLO      |      SLA      |
-|----------------|----------------------------------------------------------------------------------------------------|:-------------:|:-------------:|:-------------:|
-| **Production** | Run workloads to be used by end users.                                                             | To-be-defined | To-be-defined | To-be-defined |
-| **Weekly**     | Test and development of Radix. Radix users get to experiment with setting up and running apps.     |  Best-effort  |  Best-effort  |      No       |
-| **Nightly**    | Rapid test and integration of Radix Platform. Not to be used for workloads.                        |       No      |       No      |      No       |
-
+|    Channel     |                                           Purpose                                                  |        Termination [1]   |   Support    |     SLO [2]   |    SLA [3]    |
+|----------------|----------------------------------------------------------------------------------------------------|:------------------------:|:------------:|:-------------:|:-------------:|
+| **Production** | Run workloads to be used by end users.                                                             | TBD: Maybe quarterly     |To-be-defined | To-be-defined | To-be-defined |
+| **Weekly**     | Test and development of Radix. Radix users get to experiment with setting up and running apps.     | 2-3 weeks after creation | Best-effort  |  Best-effort  |      No       |
+| **Nightly**    | Rapid test and integration of Radix Platform. Not to be used for workloads.                        | Daily                    |      No      |       No      |      No       |
+
+> _Legend_:
+> [1] **Termination**: The lifetime before a cluster will be terminated. There will always be a newer cluster available before an old one is terminated and we will ensure application migrations to a new cluster is as smooth as possible. The reason for limiting the lifetime of a cluster is to limit the accumulation of changes to cluster state that is not documented and reproducible which can make later integrations and disaster recovery unnecessarily difficult.
+> 
 > _Abbreviations_: 
-> **SLO**: Service Level Objectives - The lower and upper bounds of stability and performance we expect to deliver. Typically measured in uptime, % of successfull requests or latency percentiles (95% of requests < 500ms) 
-> **SLA:** Service Level Agreements - Responsibility and sanctions if Service Level Objectives are not met.
+> [2] **SLO**: Service Level Objectives - The lower and upper bounds of stability and performance we expect to deliver. Typically measured in uptime, % of successfull requests or latency percentiles (95% of requests < 500ms) 
+> [3] **SLA:** Service Level Agreements - Responsibility and sanctions if Service Level Objectives are not met.
 
 
 ## Releases