Closed swarnat closed 1 year ago
Awesome, let me know when you are done, I'll test and merge!
The latest commit finish my modification to add shared workloads. Last commit was to increase version to 0.196.0. We did some internal testing of this scaling functions and don't get any problems.
An important topic was, that I was not able to upgrade from your helm chart to my one, because of a migration error in n8n itself. But because I started 7 days ago with n8n, I'm not sure, if the reason was my empty database. Maybe you, or somewhere, have a realworld database to check.
About the OCI Charts: I take single steps into the Kubernetes world. 😄 The only thing I'm currently know about OCI Charts is, they are a more general way to provide helm charts. (Maybe next-gen helm charts)
Tested this before the branch conflicts and it worked flawlessly, thanks for the work. Having a solid kubernetes HA nocode tool will make many people happy.
To save some time, here's a minimum set of values necessary for notes in case users want to test themselves.
Note: You'll need a postgreSQL instance setup ahead of time to connect:
n8n:
encryption_key: <generate and replace with encryption string>
config:
database:
type: postgresdb
postgresdb:
database: n8n
host: <postgresdb url>
port: 5432
user: <postgresdb user>
secret:
database:
postgresdb:
password: <postgresdb password>
ingress:
enabled: true
hosts:
- host: your.n8n.hostname
paths:
- /
tls:
- secretName: n8n-tls
hosts:
- your.n8n.hostname
autoscaling:
enabled: true
minReplicas: 1
maxReplicas: 2
targetCPUUtilizationPercentage: 80
scaling:
enabled: true
worker:
count: 2
concurrency: 2
webhook:
enabled: true
count: 1
redis:
host: <redis host headless>
password: <redis password>
redis:
enabled: true
global:
redis:
password: <redis password>
architecture: standalone
master:
persistence:
enabled: false
existingClaim: ""
size: 2Gi
Thanks for tuning the chart, looking forward to merge.
Happy to hear, @swarnat can you resolve the issues and I'll merge your changes!
Of course. I will solve it within next day
Conflict is solved and directly updated to latest version Update is quickly tested in our environment. We are using this "HA" setup since october and do not get any issues.
I currently play a little bit to integrate the scaling options of n8n into this helm chart.
I would be happy about users, who have the chance to test. Because my lack of experience with n8n maybe oversee something. But in my eyes a deployment to k8s don't make sense without decentralized workers/incoming webhooks and a little bit HA options
Other topics of this merge:
ToDo before Merge:
Any Feedback is welcome