When there is an error in a Helm chart, Argo saves the error message in following path:
That error message is a huge string because Helm chart and other stuff rendered into the script before executing. It was around 1 MB on a cluster I manage. Because Argo makes updates to the application definition continuously and etcd also keeps previous revisions of a key, this adds up quickly to fill the available space in etcd database. I cannot paste the whole thing here but the error message starts with
STDERR:\n Error: failed to parse /tmp/helmfile1743267544/cenkalti-app2-values-191fc62d6a: error converting YAML to JSON: yaml: line 54: could not find expected ':'\n\nCOMBINED OUTPUT:\n Error: failed to parse /tmp/helmfile2840360558/cenkalti-app2-values-191fc62d6a: error converting YAML to JSON: yaml: line 42: could not find expected ':'",
When there is an error in a Helm chart, Argo saves the error message in following path:
That error message is a huge string because Helm chart and other stuff rendered into the script before executing. It was around 1 MB on a cluster I manage. Because Argo makes updates to the application definition continuously and etcd also keeps previous revisions of a key, this adds up quickly to fill the available space in etcd database. I cannot paste the whole thing here but the error message starts with
and ends with
My suggestion is to remove
https://github.com/travisghansen/argo-cd-helmfile/blob/30b77b7a716a69d56d3f90ce340f86e6243a3ae3/src/argo-cd-helmfile.sh#L42
or comment out and add a comment to link to this issue.