Skip to content

Addon Configuration (WIP)

This page will cover the global configuration of various system addons, such as JANUS, RTMon, and IRIWM.

  • JANUS settings.
  • IRIWM settings.
  • The operator-facing configuration boundary for these addons.
  • Links to the user guide for each addon.

Procedures will be added after the operator workflow and configuration contract are verified. No enable or disable guidance is provided here.

RTMon (Real-Time Monitoring) polls this orchestrator for assigned tasks, retrieves the manifest for each corresponding service instance, and generates Grafana dashboards for the provisioned path. A single RTMon deployment may serve several orchestrators at once, so the values below must agree with that deployment’s own configuration rather than being chosen per orchestrator. For the user-facing description of the addon, see SENSE RTMon.

SettingDescription
EnabledWhether the addon is offered on this orchestrator. Disabling it does not remove dashboards that RTMon has already created.
Username for DelegationThe account that instances are automatically delegated to for RTMon access. This must be the account the RTMon deployment authenticates as, commonly sense-rtmon.
Task Assignee TagThe username, role, or group that RTMon tasks are assigned to. RTMon requests tasks by this exact assignee, so it must match the value the deployment expects, commonly rtmon.instance-manager.
Exposed DeploymentsWhich RTMon deployments are available for selection. The list is populated from RTMon deployments that have registered themselves with this orchestrator; it cannot be typed in and is empty until one has.

A registered deployment appears under the name the RTMon deployment publishes, which is derived from its Grafana folder and defaults to Real Time Mon. Because one RTMon deployment serves several orchestrators under a single name, the same value is expected on each orchestrator it serves.

Both sides of the assignee tag are configurable. The deployment asks for tasks under the tag in its own senseo_assignee setting, which defaults to rtmon.instance-manager, so a deployment can be pointed at a differently named addon without a code change. The two values still have to agree; which of them moves is an operational choice.

The per-instance options users see under the addon are declared by the RTMon deployment, not by this orchestrator, and two of them carry deployment policy that is worth knowing before a user asks about it. For the options themselves, see SENSE RTMon.

OptionDeployment policy
Dashboard RetentionThe deployment sets the default the field shows and a ceiling it will not exceed. A request above the ceiling is reduced to it rather than refused, and a default of 0 means a cancelled instance takes its dashboard with it.
The ping, traceroute, and performance testsThese are submitted by the deployment directly to each SiteRM. A deployment that cannot do so holds them off, and they are then advertised with (temporarily off) appended, so a user sees a parked option instead of selecting a test that never runs.
SymptomLikely cause
Exposed Deployments is empty and no deployment can be chosenRTMon has not completed registration. Confirm the orchestrator accepts the metadata domain RTMon publishes under. Older builds define only RTMON and reject RTMon with No enum constant SENSEMetadataDomain.RTMon.
RTMon registers and polls, but no dashboards are producedThe deployment name or Task Assignee Tag does not match what the RTMon deployment expects. Compare both against an orchestrator where the same deployment is working.
Tasks are assigned, but instance lookups failThe delegation account is missing API roles. Service instance discovery returns 403 with denied access to API due to missing roles. Grant the account the same roles it holds on a working orchestrator.
Users report that a test option is labelled (temporarily off)The deployment is holding off the tests it submits directly to a SiteRM. This is a deployment setting, not an orchestrator one, and the marker is what the deployment publishes in place of the plain label.
A user asks why their retention request was shortenedThe request exceeded the deployment’s ceiling and was reduced to it. The deployment logs the requested and the applied value.

Registration and role errors are reported by the RTMon deployment rather than surfaced in the portal, so its logs are the authoritative source when a deployment does not appear.

  • SENSE RTMon for the options a user selects per instance.