Custom TLS certificates let EdgeConnect reach hosts that don't use publicly trusted certificates. Secrets let apps and workflows send credentials without storing them. Both are optional, and both need two parts:
Where you add certificate_paths and secrets depends on how you deploy EdgeConnect:
edgeConnect.yaml file. See Deploy EdgeConnect on Docker.edge-connect-config Secret in deployment.yaml. You can't set either one through environment variables or mapped files, so mount the configuration file. See Configure EdgeConnect deployment.EdgeConnect custom resource. See EdgeConnect Operator parameters.Over HTTPS and secure WebSockets, EdgeConnect verifies a host by its certificate and trusts it only if a publicly trusted certificate authority (CA) signed it. Internal services often use a private CA or a self-signed certificate, and an intercepting proxy re-signs traffic with its own certificate. Such connections fail until EdgeConnect trusts the signing certificate.
Add the CA certificate that directly or indirectly signs the host's certificate, not the host certificate itself. EdgeConnect supports PEM (.pem, .crt, .cer) and DER (.der) files.
Add the in-container path to certificate_paths, or set the EDGE_CONNECT_CERTIFICATE_PATHS environment variable. By convention, certificates go in /etc/edge_connect_certs.
certificate_paths:- "/etc/edge_connect_certs/certificate.pem"
Mount the certificate into the container at that path.
docker run \--mount type=bind,src=${PWD}/edgeConnect.yaml,dst=/edgeConnect.yaml \-d --restart always \-v /host/path/to/certificate.pem:/etc/edge_connect_certs/certificate.pem \dynatrace/edgeconnect
If a proxy performs TLS interception, also add the proxy's own certificate to certificate_paths.
For common verification errors, see TLS certificate troubleshooting.
Apps and workflows need credentials, such as API tokens or passwords, to call services in your network. Anyone who can open or export the app can read a credential stored in it.
Put a placeholder token in the request instead, and define the real value in the EdgeConnect configuration. EdgeConnect replaces the token in query parameters, headers, and bodies immediately before it forwards the request, so the app never holds the value. EdgeConnect rejects a request that carries the token to a host outside restrict_hosts_to.
Each secret needs a generated token, a value the container reads from an environment variable or a file, and a list of allowed hosts.
Add the secrets to the EdgeConnect configuration. Configure exactly one source per secret: from_env or from_file. For the token format and all fields, see Secrets.
To generate a cryptographically secure token on Linux, run the following command. It requires openssl and the base32 utility, which is part of GNU coreutils and typically preinstalled.
echo "dt0e01.`openssl rand -out /dev/stdout 15 \| base32 \| tr '[:lower:]' '[:upper:]'`.`\openssl rand -out /dev/stdout 40 \| base32 \| tr '[:lower:]' '[:upper:]'`"
The example defines two secrets: one from an environment variable and one from a file, both restricted to hosts in restrict_hosts_to.
secrets:- name: My secrettoken: dt0e01.some-token-name.some-token-secretfrom_env: MY_SECRETrestrict_hosts_to:- dynatrace.com- name: My other secrettoken: dt0e01.another-token-name.another-token-secretfrom_file: /path/to/my/other/secretrestrict_hosts_to:- internal.example.com
Provide the values to the container. Mount the file at the from_file path.
docker run \--mount type=bind,src=${PWD}/edgeConnect.yaml,dst=/edgeConnect.yaml \-d --restart always \-e MY_SECRET=******* \-v /host/path/to/my/other/secret:/path/to/my/other/secret \dynatrace/edgeconnect
For all parameters and environment variables, see EdgeConnect reference.