Install & configure NRDOT for MySQL monitoring with Helm

You can install and configure MySQL monitoring on Kubernetes using the mysql-otel Helm chart. Optionally, the chart can also run a setup job that creates the monitoring user for you.

중요

The chart deploys the collector as a Kubernetes Deployment that reaches MySQL remotely over the network. As a result it always connects over TCP (Unix-socket monitoring isn't supported), and it reports MySQL metrics only (no host or infrastructure metrics for the machine MySQL runs on). Using mysql.*, it monitors one MySQL instance per release; to monitor several instances from a single release, use mysqlMulti.* instead (see Monitor multiple instances from one release below).

Prerequisites

A valid New Relic license key.
MySQL 5.7 or later.
Helm 3.0 or later, and kubectl configured to access your Kubernetes cluster.
Network connectivity from your Kubernetes cluster to the MySQL host (or RDS endpoint) and port:
- Self-hosted: a routed path to the MySQL host (VPN, peering, or shared network), DNS resolution if it's a hostname, and the MySQL-side firewall must allow the connection's actual source IP (which may be a NAT gateway, not the pod IP itself).
- AWS RDS: VPC peering, a transit gateway, or shared-VPC placement with the RDS instance, and the RDS security group must allow the MySQL port (3306 by default) from the cluster's egress source.
Either a monitoring user you've already created, or MySQL admin credentials so the chart's setup job can create one for you. See Configure monitoring credentials below.

Create the namespace

Create the namespace you'll install into. It must exist before you create any secrets in the next step, since helm install --create-namespace only creates the namespace during the final install step:

bash

$kubectl create namespace newrelic

팁

Set it as the default namespace for your current kubectl context so you don't need to pass -n on every command below:

bash

$kubectl config set-context --current --namespace=newrelic

Configure monitoring credentials

Decide whether you want the chart's setup job to create the MySQL monitoring user for you, and if not, how you'll supply its credentials:

With this method there's no setup job, so nothing here creates the monitoring user for you. Before installing the chart, create it yourself using the Configure database user step from self-hosted MySQL or MySQL on RDS, or run the equivalent SQL as an admin user:

CREATE USER IF NOT EXISTS 'newrelic'@'%' IDENTIFIED BY '<password>';
GRANT SELECT ON performance_schema.* TO 'newrelic'@'%';
FLUSH PRIVILEGES;

You'll set the monitoring username and password directly in values.yaml in the next step.

With this method there's also no setup job. Create the monitoring user yourself first, exactly as in the plain-credentials method above. The only difference is where its credentials live: instead of plaintext in values.yaml, store them in a Kubernetes secret:

bash

$kubectl create secret generic mysql-monitor-creds \
>  --from-literal=username=<MONITORING_USERNAME> \
>  --from-literal=password='<MONITORING_PASSWORD>' \
>  -n newrelic

With this method, the chart's setup job creates the monitoring user for you using MySQL admin credentials. Both the monitoring and admin credentials must be provided as Kubernetes secrets: the setup job never accepts plaintext admin credentials, and enabling it requires secret-based monitoring credentials too.

Create a secret with the monitoring username and password. This is the identity the setup job creates and the collector connects with:

bash

$kubectl create secret generic mysql-monitor-creds \
>  --from-literal=username=<MONITORING_USERNAME> \
>  --from-literal=password='<MONITORING_PASSWORD>' \
>  -n newrelic

The setup job only accepts admin credentials via a Kubernetes secret (no plaintext username/password in values.yaml). Since this credential can run CREATE USER and grant broad performance_schema access, create it as its own secret. Use the root account for self-hosted MySQL, or the RDS master user for MySQL on AWS RDS:
bash
```
$kubectl create secret generic mysql-admin-creds \
>  --from-literal=username=<ADMIN_USERNAME> \
>  --from-literal=password='<ADMIN_PASSWORD>' \
>  -n newrelic
```

Configure the Helm values

Set mysql.topology to match your environment:

Value	Description
`self-hosted`	Self-hosted MySQL, reached over the network.
`rds`	MySQL or Aurora on AWS RDS.

Use the values.yaml that matches the credential method you chose in the previous step:

licenseKey: <YOUR_LICENSE_KEY>
otlpEndpoint: otlp.nr-data.net:4317

mysql:
  topology: <self-hosted|rds>
  server: <YOUR_DB_HOST>
  port: <YOUR_DB_PORT>
  username: <YOUR_MONITORING_USERNAME>
  password: <YOUR_MONITORING_PASSWORD>

# No setup job runs, so create the monitoring user yourself
# before installing the chart.
setupJob:
  enabled: false

licenseKey: <YOUR_LICENSE_KEY>
otlpEndpoint: otlp.nr-data.net:4317

mysql:
  topology: <self-hosted|rds>
  server: <YOUR_DB_HOST>
  port: <YOUR_DB_PORT>
  existingSecret: mysql-monitor-creds

# No setup job runs, so the monitoring user referenced by
# this secret must already exist - create it yourself
# before installing the chart.
setupJob:
  enabled: false

licenseKey: <YOUR_LICENSE_KEY>
otlpEndpoint: otlp.nr-data.net:4317

mysql:
  topology: <self-hosted|rds>
  server: <YOUR_DB_HOST>
  port: <YOUR_DB_PORT>
  existingSecret: mysql-monitor-creds

setupJob:
  enabled: true
  mysqlAdmin:
    existingSecret: mysql-admin-creds
  # -- Also grants UPDATE on performance_schema.setup_consumers,
  # needed for wait-time data. Optional.
  enableWaitTimeMetrics: false

Parameter	Description
`licenseKey`	Your New Relic license key.
`otlpEndpoint`	Your region's OTLP/gRPC endpoint, as a bare `host:port` with no scheme: `otlp.nr-data.net:4317` (US) or `otlp.eu01.nr-data.net:4317` (EU). For more information, refer to New Relic OTLP endpoints documentation.
`mysql.server` / `mysql.port`	Your MySQL host (or RDS endpoint) and port, reachable from the cluster.
`mysql.database`	Optional. Restrict monitoring to one database. Leave empty to monitor all databases.
`additionalReceiverConfig.tls.*`	Optional. The chart doesn't declare a `tls` block and has no `mysql.tls.` values. The receiver's own defaults apply (`insecure: false`, `insecure_skip_verify: false`), which require an encrypted connection with certificate validation. Override them through `additionalReceiverConfig`: set `additionalReceiverConfig.tls.insecure_skip_verify=true` to skip certificate validation (for testing with self-signed certificates only, since it weakens the connection's security), or `additionalReceiverConfig.tls.ca_file` to point at a CA bundle mounted into the collector container. An Amazon RDS instance with Require SSL/TLS enforcement needs `ca_file` set to Amazon's RDS CA bundle*. The chart deliberately ships no default bundle, since a hardcoded bundle risks going stale as Amazon rotates CAs.
`setupJob.image.repository` / `setupJob.image.tag`	The `mysql` CLI image used by the setup job. Defaults to the official, actively maintained `mysql:8.4`, so enabling the setup job needs no extra image flags. Override only if you need a different version.
`setupJob.enableWaitTimeMetrics`	Optional. Also grants `UPDATE` on `performance_schema.setup_consumers`, which is needed for wait-time data.

팁

See the chart's values.yaml for all available configuration options, including mysql.collectionInterval, mysql.explainMode, the statementEvents/querySampleCollection/topQueryCollection blocks, and the additionalReceiverConfig escape hatch.

(Optional) Monitor multiple instances from one release

Instead of mysql.*, set mysqlMulti.enabled: true to monitor several MySQL instances (self-hosted or RDS) from a single collector pod in one release. The two modes are mutually exclusive: don't set mysql.server and mysqlMulti.enabled: true in the same release.

Multi-instance mode has no plain-username/password path: every instance needs its own Kubernetes secret with monitoring credentials. Decide whether you also want the setup job to create each monitoring user for you:

Create each instance's monitoring-user secret yourself before installing (repeat per instance, matching the name you give it in values.yaml):

bash

$kubectl create secret generic db1-monitor-creds \
>  --from-literal=username=<MONITORING_USERNAME> \
>  --from-literal=password='<MONITORING_PASSWORD>' \
>  -n newrelic
$# Repeat for every instance (db2-monitor-creds, ...)

licenseKey: <YOUR_LICENSE_KEY>
otlpEndpoint: otlp.nr-data.net:4317

# EDIT the two example entries below (or add more) to match
# your real instances - each `name` must be unique within
# this release and match the Secret you created for it.
mysqlMulti:
  enabled: true
  topology: <self-hosted|rds>
  databases:
    - name: db1
      server: <YOUR_DB_1_HOST>
      existingSecret: db1-monitor-creds
    - name: db2
      server: <YOUR_DB_2_HOST>
      existingSecret: db2-monitor-creds

# No admin credential is configured here since the setup
# job is disabled - create each instance's monitoring user
# yourself first.
setupJob:
  enabled: false

Each instance needs both its own monitoring-user secret and its own admin-credentials secret (repeat per instance):

bash

$kubectl create secret generic db1-monitor-creds \
>  --from-literal=username=<MONITORING_USERNAME> \
>  --from-literal=password='<MONITORING_PASSWORD>' \
>  -n newrelic
>kubectl create secret generic db1-admin-creds \
>  --from-literal=username=<ADMIN_USERNAME> \
>  --from-literal=password='<ADMIN_PASSWORD>' \
>  -n newrelic
$# Repeat for every instance (db2-monitor-creds, db2-admin-creds, ...)

licenseKey: <YOUR_LICENSE_KEY>
otlpEndpoint: otlp.nr-data.net:4317

# EDIT the two example entries below (or add more) to match
# your real instances - each `name` must be unique within
# this release and match the Secrets you created for it.
mysqlMulti:
  enabled: true
  topology: <self-hosted|rds>
  databases:
    - name: db1
      server: <YOUR_DB_1_HOST>
      existingSecret: db1-monitor-creds
      mysqlAdmin:
        existingSecret: db1-admin-creds
    - name: db2
      server: <YOUR_DB_2_HOST>
      existingSecret: db2-monitor-creds
      mysqlAdmin:
        existingSecret: db2-admin-creds

setupJob:
  enabled: true

Parameter	Description
`mysqlMulti.enabled`	Set to `true` to monitor multiple instances from one release instead of `mysql.*`. Defaults to `false`.
`mysqlMulti.topology`	`self-hosted` or `rds`, same meaning as `mysql.topology`, shared by every instance in this release.
`mysqlMulti.databases`	List of instances to monitor. Each entry requires a unique `name`, `server` (`port` defaults to `3306`), and `existingSecret`; `database` optionally restricts monitoring to one database (same meaning as `mysql.database`); add `mysqlAdmin.existingSecret` per entry only when `setupJob.enabled` is `true`.

팁

Scrape settings (collection interval, TLS, statement events, query sample collection, top query collection) apply identically to every instance in mysqlMulti.databases. There's no per-instance override. Use additionalReceiverConfig for anything that needs to differ. The setup job runs once per instance (<release>-setup-<name>) rather than once per release.

Install the Helm chart

Add the New Relic Helm repository:

bash

$helm repo add newrelic https://helm-charts.newrelic.com
$helm repo update

Install the chart using your values.yaml file:

bash

$helm upgrade --install mysql-otel newrelic/mysql-otel \
>  -n newrelic \
>  --create-namespace \
>  -f values.yaml

Verify the installation

Check that the setup job completed (if enabled) and the collector pod is running:
bash
```
$kubectl get jobs,pods -n newrelic --watch
```

Run this query in the query builder to confirm data is arriving:

SELECT count(*) FROM Metric
WHERE metricName LIKE 'mysql.%'
AND instrumentation.provider = 'opentelemetry'
SINCE 10 minutes ago

Find and use your data

Once your data is being collected, you can access comprehensive MySQL database monitoring through the New Relic UI.

To find your MySQL database entity in New Relic:

Go to one.newrelic.com > All capabilities > Databases.
From the Entity type dropdown, select MySQL instance, then click Apply.
Select your MySQL database from the list of entities.

Introduction to MySQL monitoring with NRDOT

Learn about all the available installation methods for MySQL monitoring with New Relic.

CLI install

Learn how to install and configure MySQL monitoring with a single New Relic CLI command.

Ansible install

Learn how to install and configure MySQL monitoring at scale with the newrelic.newrelic_install Ansible role.

Chef install

Learn how to install and configure MySQL monitoring at scale with the newrelic-install Chef cookbook.

Install & configure NRDOT for MySQL monitoring with Helm

중요

Prerequisites

Create the namespace

팁

Configure monitoring credentials

No setup job: plain username and password

No setup job: secret-based credentials

Automated setup job

Configure the Helm values

No setup job: plain username and password

No setup job: secret-based credentials

Automated setup job

팁

(Optional) Monitor multiple instances from one release

No setup job

Automated setup job

팁

Install the Helm chart

Verify the installation

Find and use your data

Introduction to MySQL monitoring with NRDOT

CLI install

Ansible install

Chef install

Install & configure NRDOT for MySQL monitoring with Helm

중요

Prerequisites .css-21sua1{background:none;border:none;width:0;padding:0;}

Create the namespace

팁

Configure monitoring credentials

No setup job: secret-based credentials

Automated setup job

Configure the Helm values

No setup job: plain username and password

No setup job: secret-based credentials

Automated setup job

팁

(Optional) Monitor multiple instances from one release

No setup job

Automated setup job

팁

Install the Helm chart

Verify the installation

Find and use your data

Related documentation

Introduction to MySQL monitoring with NRDOT

CLI install

Ansible install

Chef install

Prerequisites