gitops-engine
gitops-engine copied to clipboard
Democratizing GitOps
This change addresses the performance issue existing in the cluster cache described in https://github.com/argoproj/argo-cd/issues/18838. `kube-apiserver` logs for the Pods resource (supposed super-low latency logged for the WATCH requests is due...
See parent issue for details: https://github.com/argoproj/gitops-engine/issues/614 This PR: - When calling gitops-engine, you will now need to explicitly define the temporary path to write kubectl manifests to: - Add new...
At present, gitops-engine (and consumers of gitops-engine API) use `/dev/shm` to store temporary files that are passed to kubectl, such as manifests and cluster credentials. The temp directory logic can...
If loadInitialState returns an error, it is possible to still call watchEvents and loadInitialState again when the resource type in error is not restricted.
closes #459 Based on the testing matrix and the k8s versions in the go.mod, it should be safe to remove the PSP from our kinds list. https://github.com/argoproj/gitops-engine/blob/a22b34675f57696d55cb9d4abf99f7b07a166654/go.mod#L24 https://argo-cd.readthedocs.io/en/stable/operator-manual/installation/#tested-versions
argo-cd issue https://github.com/argoproj/argo-cd/issues/18857
See the issue https://github.com/argoproj/argo-cd/issues/18500 and investigation there. For new config maps (and potentially other resources) they can already be created rather quickly, but since resources in `syncContext` are based resources...
Problem statement is in https://github.com/argoproj/argo-cd/issues/8172#issuecomment-2277585238 The [IterateHierrchyV2](https://github.com/argoproj/gitops-engine/commit/6b2984ebc47085852a7b63a0fd0b73c52e986217) significantly improved performance, getting us ~90% of the way there. **But on huge clusters, we still have significant lock contention.** The fix in...
Currently the health check return healthy if the first conditions is healthy. This fix, first verify all conditions are healthy before returning healthy, allowing to catch when the first conditions...
If resources have the `argocd.argoproj.io/sync-options: Force=true,Replace=true` annotation, they are recreated twice: * when Argo CD performs dry-run * when Argo CD performs the actual sync This behavior can be observed...