Skip to main content

System for deploying certificates from Hashicorp Vault server

Project description



Deploy SSL certificates from HashiCorp's Vault secret server Script is able to deploy certificates from KV store of Vault or when you use issue version of script it use PKI secret storage.

As auth method is used Approle, you need role and secret id deployed to server from different systems/locations. More about this in usecase section.

On the first look, it may be little bit strange combination of config file and cli options. You can combine them in different ways to support various types of deployments to meet the basic security concepts.

Why do I need Vault Server ?

We are using Let's Encrypt for almost all of our SSL/TLS certificates. We also have complex infrastructure so we have to retrieve certificates in central place and then we distribute them into datacenters, clouds or any other applications.

How deploy work ?

It deploy certificates to specified directory and create two directories certs and private.

  • certs has mode 0644
  • private keys has mode 0640
  • it deploys all secret content from vault, keys as files with suitable extension .


Python PyPI

pip install vault-cert-deploy


Manual installation

git clone
cd vault-certificate-deploy
sudo python ./ install 

In the end

vault-cert-deploy --help

Example configuration

Can be found in config.example.

Role and Secret id can be passed from script arguments. You could combine -n and --cert-list parameters.

In vault section of configuration it is possible to set mount_point of secret in Vault. By default it is cert. You could also change this option in arguments

Vault Configuration

Script uses Approle auth.

First enable AppRole auth if it's not

vault auth enable approle

You have to create your policy first. Use Vault documentation around policies and then continue here.

Example policy with basic medium security can be

# Cert Deploy Policy
# Give ability to
# - read all certificates
# - don't permit list certificates
path "/certs/*" {
  capabilities = ["read"]

Configure your role

vault write auth/approle/role/my-role \
secret_id_ttl=0 \
token_num_uses=0 \
token_ttl=20m \
token_max_ttl=30m \

Retrieve your approle ID

vault read auth/approle/role/my-role/role-id

Get secret ID (onetime operation)

vault write -f auth/approle/role/my-role/secret-id

Use Cases

It is important to don't have role-id and secret-id together in one repository or configuration management.


I deploy my servers with installer which create file /etc/vault_role_id which contain role-id of the approle.

Then I have Puppet Configuration management which deploy this system with all files and secret-id in configuration file (/etc/vault-deploy/config.conf).

Puppet create also file with certs/secret names /etc/ssl-deploy-certs.conf

then you can run deploy like this: -c /etc/vault-deploy/config.conf \
  --cert-list /etc/ssl-deploy-certs.conf \
  --role-id $(cat /etc/vault_role_id)

Why ?

I store Puppet configuration in Git, and therefore I have not role-id and secret-id together in my repository. I choose to deploy secret-id with puppet because when need to rotate secret-id it is automaticly deployed by puppet to infrastructure.

What is issue version of the script ?

Issue version of the command or script uses different Secret Storage Engine. It uses PKI which gives you ability to create your own CA or Intermediate CA. Vault handle both certs generation and issuing.

You have to specify PKI mount point with --vault-pki option. This pki mount_point is used as subdirectory of storage path in your config file. In this subdirectory we create same structure certs and private like in other version of the script.

What is difference in function ?

Issue command check if certificates you define exists, and it check their expiration time defined by --cert-min-ttl option.

It basicaly means it generates and issue certificates for you, if they not exist, or if they are close to expire. It is great automation capability in combination with Configuration Management systems. You don't have to take care of the certificates anymore.

If certificates you define exists and are valid script just do nothing.


Create certificate server1.domin.intra on PKI mounted in pki mount point of vault. If you want to issue new certificate, you have to issue it against some role. In our case this role is test.

More information about PKI roles in documentation.

vault-certificate-issue-deploy --vault-pki pki -n server1.domain.intra --cert-role test

If we need some subject alternative name you can define it as --cert-extra-options

vault-certificate-issue-deploy --vault-pki pki -n server1.domain.intra --cert-role test --cert-extra-options "alt_names=console.domain.intra,console1.domain.intra,admin.domain.intra"

Result of this can be something like this

 X509v3 Subject Alternative Name: 
     DNS:console.domain.intra, DNS:console1.domain.intra, DNS:admin.domain.intra

Security Best Practices

  • Never store your role-id and secret-id in your repository together
  • Deploy secret-id in way it's quick and easy for you to rotate/change
  • In production always use verify_tls=yes
  • when deploy secret-id and role-id in files/config, always set correct permissions (eg. 0400, 0600)
  • in vault set policy to your approle only for read capability, it's enough
  • for highest security set individual approle for every server and set individual policy for every server and certificate

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for vault-certificate-deploy, version 1.2.1
Filename, size File type Python version Upload date Hashes
Filename, size vault_certificate_deploy-1.2.1-py2-none-any.whl (27.7 kB) File type Wheel Python version py2 Upload date Hashes View
Filename, size vault_certificate_deploy-1.2.1-py3-none-any.whl (27.7 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size vault-certificate-deploy-1.2.1.tar.gz (14.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page