Configuring Traefik timeout values for large file uploads and downloads

When uploading or downloading large files, the default timeout value for readTimeout, writeTimeout and idleTimeout errors in Traefik proxy might not be sufficient. You can set the Traefik proxy timeout to a new duration instead of the default value.

Back up the ConfigMap.

kubectl -n [***NAMESPACE***] get cm traefik-config -o yaml > /tmp/traefik-config.backup.yaml

Edit the Traefik ConfigMap.

kubectl edit configmap traefik-config -n [***NAMESPACE***]

You can optionally add or update the timeout value fields to manually specify the timeout values in the data.traefik.toml section of the ConfigMap file in Cloudera AI on premises 1.5.5 SP2 and higher releases.
```
[entryPoints.web.transport.respondingTimeouts]
readTimeout = [***TIMEOUT VALUE***]
  writeTimeout = [***TIMEOUT VALUE***]
  idleTimeout = [***TIMEOUT VALUE***]

[entryPoints.websecure.transport.respondingTimeouts]
  readTimeout = [***TIMEOUT VALUE***]
  writeTimeout = [***TIMEOUT VALUE***]
  idleTimeout = [***TIMEOUT VALUE***]
```
note
In Cloudera AI on premises 1.5.5 SP1 and lower releases you need to add the above specified timeout fields to the data.traefik.toml file. The default timeout value for readTimeout, writeTimeout and idleTimeout errors is 60 seconds in the Traefik proxy in Cloudera AI on premises 1.5.5 SP1 and lower releases. You can also configure the default timeout to different values if needed.

Restart the Traefik proxy to ensure it applies the changes.

kubectl -n [***NAMESPACE***] rollout restart deployment/ingress-controller

Verify that the changes are implemented.

kubectl -n [***NAMESPACE***] get cm traefik-config -o jsonpath='{.data.traefik\.toml}' | grep -A3 respondingTimeouts

To perform a rollback, use the following commands:

kubectl -n [***NAMESPACE***] apply -f /tmp/traefik-config.backup.yaml
kubectl -n [***NAMESPACE***] rollout restart deployment/ingress-controller