Skip to main content

Telemetry

Artemis reports installation configuration to help the maintainer team understand deployments, support their configurations, and plan improvements. Set artemis.telemetry.enabled: false to disable reporting.

Startup reporting​

Only the node with both core and scheduling profiles reports telemetry. Configure exactly one scheduling node per installation. It sends one report ten minutes after the application becomes ready, allowing other nodes and build agents to connect. Counts are collected when the report is sent. Restarting another core node or a build agent does not produce a report, and shutting down the scheduling node before the delay expires cancels its pending report.

Development servers and servers configured with info.testServer: true do not send telemetry. Artemis does not detect test servers on its own, so set the flag explicitly on every test installation.

The collector identifies each installation by its URL and keeps a separate startup history. Repeated delivery of the same startup identifier does not add duplicate entries. The latest startup supplies the current installation snapshot, even if an older report arrives later.

Required configuration​

Every production core node refuses to start unless these properties contain meaningful values, independently of telemetry settings. Development servers, test servers (info.testServer: true), and dedicated build agents start without them:

  • info.operatorName: the organization operating Artemis.
  • info.operatorAdminName: the administrator's name.
  • info.universityName: the university, school, or institution using Artemis.

Empty, whitespace-only, and known template values such as Admin, Some Artemis Operator, Your University, and angle-bracket placeholders are rejected. The error lists the settings to correct. Disabling telemetry does not bypass this startup check. The About page displays the values that are set. The university and operator are independent; neither is inferred from the other.

artemis.telemetry.sendAdminDetails: false omits the administrator's name and contact email from transmission. It does not bypass local configuration validation or hide the administrator name from the About page.

artemis:
telemetry:
enabled: true
sendAdminDetails: false
destination: https://telemetry.artemis.cit.tum.de

info:
testServer: false
operatorName: <organization operating Artemis>
universityName: <university or institution>
operatorAdminName: <administrator name>
contact: <administrator email address>

Reported information​

FieldMeaning
version, serverUrl, operator, universityNameInstallation version, URL, operator and institution
profilesActive Spring profiles on the scheduling node
ciSystem, versionControlSystem, distributedDataProvider, websocketBrokerThe CI system, version control system and middleware of the installation, see below
moduleFeaturesEnabled module features resolved from Artemis configuration; these are not measurements of user activity
numberOfNodesCore Artemis nodes with recent registry heartbeats
buildAgentCountConnected build agents, including paused agents and core nodes that also build
isMultiNodeWhether more than one core node is connected
dataSourceThe database in use, postgresql or mysql. Any JDBC URL that does not start with jdbc:mysql is reported as postgresql
isProductionInstance, isTestServer, isLocalLLMDeploymentEnabledDeployment configuration
startupId, startedAtIdentifier and start time of this startup of the scheduling node's application
adminName, contactAdministrator details, only when sendAdminDetails is enabled

Dedicated build agents are counted separately from core nodes. A node with both roles contributes to both counts. Missing connectivity information is omitted rather than reported as zero; without LocalCI, the build-agent count is zero. Ten minutes allows discovery to settle, but the report describes the observed topology at that time, not a guarantee that all configured nodes are healthy. The collector records its own receipt time separately from startedAt.

CI system and middleware​

Newer Artemis versions report which systems an installation runs. The maintainers use these values to decide which integrations to keep supporting, for example before removing Jenkins or MySQL in Artemis 11.

FieldValuesDerived from
ciSystemlocalci, jenkins, noneThe localci and jenkins profiles of the scheduling node. none means neither is active, for example an installation without programming exercises.
versionControlSystemlocalvc, noneThe localvc profile of the scheduling node
distributedDataProviderhazelcast, redis, localartemis.distributed-data.provider, or the superseded artemis.continuous-integration.data-store. Hazelcast is the default.
websocketBrokerexternal, simpleexternal when spring.websocket.broker.addresses contains at least one valid address, so that websocket messages are relayed to a broker such as ActiveMQ Artemis. simple is the in-memory broker of a single node.

dataSource reports postgresql or mysql and buildAgentCount the number of connected build agents, which is 0 without LocalCI. Reports of older Artemis versions do not contain the four fields above, and the collector keeps them unknown.

Only the scheduling node reports, so the values describe that node. Dedicated build agent nodes do not report; they are visible only through buildAgentCount.

Search documentation