This example demonstrates how to run the Cloud SQL Auth Proxy with PgBouncer on Kubernetes as a service. It assumes you have already successfully completed all the steps in Using the Cloud SQL Auth Proxy on Kubernetes.
In this example, you will deploy PgBouncer with the Cloud SQL Auth Proxy as a sidecar, in addition to configuring encryption between the application and PgBouncer.
Running PgBouncer with the Cloud SQL Auth Proxy may pose a significant operational burden and should be undertaken with caution given the attendant complexity.
In general, we recommend running the proxy as a sidecar to your application because it is simple, there is less overhead, it is secure out of the box, and there is less latency involved.
However, the service pattern is useful when you are at very large scale, when you clearly need a database connection pooler, and when you are running into SQL Admin API quota problems.
Before we deploy PgBouncer with the Cloud SQL Auth Proxy, there are three initial steps to take.
First, you will need to generate certificates to encrypt the connection between the application and PgBouncer. We recommend using CFSSL to handle certificate generation. Note: this example uses self-signed certificates. In some cases, using a certificate signed by a public certificate authority may be preferred. Alternatively, Kubernetes includes an API for issuing certificates. See the documentation on certificates for more details.
The certificate signing request is encoded as JSON in
ca_csr.json
for the certificate authority and in
server_csr.json
for the "server," here PgBouncer.
First, we initialize our certificate authority.
# This step produces ca-key.pem (the CA private key)
# and ca.pem (the CA certificate).
cfssl genkey -initca ca_csr.json | cfssljson -bare ca
Next, we generate a public and private key for the server. These will be what we will use to encrypt traffic from the application to PgBouncer.
# This step produces server-key.pem (the server private key)
# and server.pem (the server certicate).
cfssl gencert -ca cert -ca-key key server_csr.json | cfssljson -bare server
Second, with all the necessary certificates generated, we will save them as secrets:
# First the CA cert
kubectl create secret tls <YOUR-CA-SECRET> --key="ca-key.pem" --cert="ca.pem"
# Next the server cert
kubectl create secret tls <YOUR-SERVER-CERT-SECRET> --key="server-key.pem" \
--cert="server.pem"
Third, we will containerize PgBouncer. Some users may prefer to containerize
PgBouncer themselves. For this example, we will make use of an open source
container, edoburu/pgbouncer. One nice benefit of edoburu/pgbouncer
is that it will generate all the PgBouncer configuration based on environment
variables passed to the container.
With PgBouncer containerized, we will now create a deployment with PgBouncer and the proxy as a sidecar.
First, we mount our CA certificate and server certificate and private key,
renaming the certificate secrets to cert.pem
and server private key to
key.pem
:
volumes:
- name: cacert
secret:
secretName: <YOUR-CA-SECRET>
items:
- key: tls.crt
path: cert.pem
- name: servercert
secret:
secretName: <YOUR-SERVER-CERT-SECRET>
items:
- key: tls.crt
path: cert.pem
- key: tls.key
path: key.pem
Next, we specify volume mounts in our PgBouncer container where the secrets will be stored:
- name: pgbouncer
image: <PG-BOUNCER-CONTAINER>
ports:
- containerPort: 5432
volumeMounts:
- name: cacert
mountPath: "/etc/ca"
readOnly: true
- name: servercert
mountPath: "/etc/server"
readOnly: true
Then we configure PgBouncer through environment variables. Note: we use 5431 for
DB_PORT
to leave 5432 available.
env:
- name: DB_HOST
value: "127.0.0.1"
- name: DB_USER
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: username
- name: DB_PASSWORD
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: password
- name: DB_NAME
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: database
- name: DB_PORT
value: "5431"
- name: CLIENT_TLS_SSLMODE
value: "require"
- name: CLIENT_TLS_CA_FILE
value: "/etc/ca/cert.pem"
- name: CLIENT_TLS_KEY_FILE
value: "/etc/server/key.pem"
- name: CLIENT_TLS_CERT_FILE
value: "/etc/server/cert.pem"
For the PgBouncer deployment, we add the proxy as a sidecar, starting it on port 5431:
- name: cloud-sql-proxy
image: gcr.io/cloud-sql-connectors/cloud-sql-proxy:2.0.0 # make sure to use the latest version
args:
# Replace DB_PORT with the port the proxy should listen on
- "--port=<DB_PORT>"
- "<INSTANCE_CONNECTION_NAME>"
securityContext:
runAsNonRoot: true
Next, we create a PgBouncer service, listening on port 5342:
apiVersion: v1
kind: Service
metadata:
name: <YOUR-SERVICE-NAME>
spec:
selector:
app: <YOUR-APPLICATION-NAME>
ports:
- protocol: TCP
port: 5432
targetPort: 5432
With the PgBouncer service and deployment done, we are ready to point our application at it.
First, we configure a volume for the CA certificate, mapping the file name to
cert.pem
.
volumes:
- name: cacert
secret:
secretName: <YOUR-CA-CERT>
items:
- key: tls.crt
path: cert.pem
Next, we mount the volume within the application container:
volumeMounts:
- name: cacert
mountPath: "/etc/ca"
readOnly: true
Then, we configure environment variables for connecting to the database, this
time including a CA_CERT
:
env:
- name: DB_HOST
value: "<YOUR-SERVICE-NAME>.default.svc.cluster.local" # using the "default" namespace
- name: DB_USER
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: username
- name: DB_PASS
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: password
- name: DB_NAME
valueFrom:
secretKeyRef:
name: <YOUR-DB-SECRET>
key: database
- name: DB_PORT
value: "5432"
- name: CA_CERT
value: "/etc/ca/cert.pem"
Note: now the DB_HOST
value uses an internal DNS record pointing at the
PgBouncer service.
Finally, when configuring a database connection string, the application must provide the additional properties:
sslmode
must be set to at leastverify-ca
sslrootcert
must set to the environment variableCA_CERT