-
### What happened?
We operate a large cluster for data processing, we use argo for job orchestration, there different kind of jobs, some of them may run longer time than others, when pods number ex…
-
### Description of Issue
A multi-step orchestration started by a reactor is not soft-killed as expected by the `state.soft_kill` runner.
### Setup
Set up a reactor to trigger a long-running o…
-
**Issue/Feature Description:**
Review and improvement of the Left Side Main Menu Items for docs.sodafoundation.io
**ALL DETAILS HERE:**
- [ ] Introduction (New) -- Give intro about SODA Foundat…
-
**Description**
We observed an issue with the orchestrator, when it has to manage long states.
In details, the orchestrator runs a state that takes about 4 hours to complete (it performs a massive d…
-
Develop deployment scripts and templates
• recommend a deployment solution (e.g. GitHub Actions, Azure Pipelines, Jenkins, CircleCI, etc.)
• design and implement Infrastructure as code (ARM, Terra…
-
We have recently encountered an increasing number of cases where the `WaitForOrchestrationAsync` method returns null for an orchestration instance that is already in the `Completed` status. This issue…
-
### Description of Issue
I am spinning up a vm on digitalocean. The vm is to have a minion and master process running on it upon completed provisioning. The minion options populate but not the master…
-
## Problem
We want to make sure that datadog-service works well with 0.20 version of Keptn since this version is going to be the baseline for Keptn 1.0 LTS
## TODO
- [x] add 0.20 in the integrati…
-
How about some API for the executor? Some information that would be exposed read/write though the API:
1. ACE's collection of PODs
2. POD run state, I/O, and filesystem
3. ACE's availability (i.e. hea…
-
When DD testing has been done in the NHS Digital instance, training will be required to other DMDS staff.
DMDS have discussed with Data Products the set up of training environment.
Cath and Angie w…