Ondat Telemetry

Ondat collects telemetry and error reports from Ondat clusters via two different methods for two different purposes.

Telemetry

  • Telemetry is made up of the DNS query and crash reports sent to sentry.io once per day.
  • Error reporting is the sentry.io crash dump reporting.

sentry.io

Ondat sends crash reports and information about the Ondat cluster to sentry.io. This information helps our developers monitor and fix crashes. Information is sent to sentry.io once per day or when a process inside the Ondat container crashes.

  • The once per day report includes the cluster version, node counts and license information.
  • The crash report contains the signal that triggered the shutdown (e.g. SIGSEGV), the exit code and whether or not the crash generated a core dump.

All Ondat clusters with a routable connection to the internet will send crash reports to sentry.io over tcp/443. Ondat respects environment variables that ProxyFromEnvironment uses.

An exhaustive list of information included in the once per day report is below:

  • API version
  • Cluster ID
  • CPU architechture
  • Go version
  • Healthy volume count
  • Logging level
  • Logging user
  • License type
  • Namespace count
  • Node count
  • Pool count
  • OS type
  • Rule count
  • Server name
  • Sentry version
  • Ondat build information (CI pipeline reference)
  • Ondat version
  • Suspect volume count
  • Total volume size
  • Volume count
  • Volume degraded count
  • Volume offline count
  • Volume syncing count

An exhaustive list of information included in the crash report is below:

  • API version
  • Cluster ID
  • CPU architechture
  • Crashed component name
  • Error level
  • Error message
  • Exceptions
  • Exit code
  • Datetime of crash
  • Go version
  • Kernel signal
  • OS type
  • License type
  • Logging user
  • Server name
  • Stacktrace
  • Ondat build information (CI pipeline reference)
  • Ondat version
  • Whether a core dump occured

DNS Query

Ondat will also send anonymized node ids, cluster id and Ondat version information to Ondat using a DNS query. The information that we send in the query is encoded as well as being anonymized. This query allows us to inform Cluster admins when Ondat upgrades are available in the Ondat GUI and in the logs.

The DNS query includes:

  • Anonymized Ondat Cluster ID
  • Anonymized Ondat node ID
  • Ondat version number

If k8sDistro is set then the Kubernetes version and Kubernetes distribution will also be reported. This information helps us direct focus onto the most relevant platforms.

Disable Telemetry

It is possible to disable telemetry using the CLI, API, environment variables or the Ondat Cluster Spec.

API

The example below shows how you can use the /v1/telemetry API endpoint to disable or enable telemetry.

$ curl -X "PUT" "http://127.0.0.1:5705/v1/telemetry"        \
         -H 'Content-Type: application/json; charset=utf-8' \
         -u 'storageos:storageos'                           \
         -d $'{
             "telemetryEnabled": true,
             "reportErrors": false
         }'

Ondat Cluster Spec

Disable telemetry explicitly through the configurable spec parameters of the StorageOSCluster custom resource.

Environment Variables

You can use the following environmental variables to disable or enable telemetry.

DISABLE_TELEMETRY       # Disable the DNS query and once per day sentry.io reporting
DISABLE_ERROR_REPORTING # Disable sentry.io crash reports

You can find more information about Ondat environment variables here.