IBM Turbonomic ARM v8.x Administrator - Professional flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Supply chain layers

The supply chain is a layered graph where supply entities such as hosts, datastores, clusters, and storage provide resources to demand entities such as VMs, pods, containers, and applications.

Market analytics engine

The decision core that computes actions by resolving contention for shared resources among supply chain entities, rather than by comparing metrics to static thresholds.

Multi-node deployment

An installation topology where platform services, the database, and analytics run on separate nodes for scale and availability, requiring an external supported database.

Installation prerequisites

Before installing, confirm supported OS version, bidirectional DNS resolution, synchronized time via NTP, and required port access between nodes and targets.

License capacity model

Licensing is capacity-based per entity type, so the license file defines covered editions and entity types, and usage is measured against licensed capacity for compliance.

Action mode default

The platform-wide default action mode, Recommend, Manual, or Automatic, applies where no more specific policy overrides it for the scoped entities.

Role-based access

Administrator, scoped site administrator, and read-only roles determine whether a user may execute actions or only view entities, actions, and reports.

SSO and LDAP mapping

Directory groups from LDAP/AD or SAML/OIDC single sign-on are mapped to platform roles, so membership changes in the directory change effective permissions.

Target credential validation

Each target's credentials are validated against the endpoint; a failed validation blocks discovery and automation for that target until corrected.

Read vs action credentials

Monitoring credentials only permit discovery, while automation credentials with higher privileges are required for the platform to execute actions on the target.

Stitching

Stitching correlates entities reported by different targets, such as a VM, its datastore, and the backing storage array, so shared resources are modeled once.

Target groups

Targets can be organized into groups so that policies and scopes apply to many targets at once instead of being configured individually.

Policy precedence

The effective policy is the most specific applicable one: entity-level settings override group-level policies, which override global platform defaults.

Placement policy

Placement policies constrain which clusters or hosts a workload may move to, including pinning and anti-affinity, limiting the actions the engine may propose.

Scaling policy

Scaling policies define whether a workload may scale and set bounds such as minimum and maximum replicas or VM size limits.

Scheduling policy

Scheduling policies define time windows during which automation is permitted or suppressed, letting administrators block changes during business-critical periods.

Resize action

A resize action changes an entity's allocated capacity up or down when sustained utilization falls outside the acceptable range set by headroom constraints.

Move action

A move action migrates a workload to another host or cluster when the current host is over-utilized and a destination can absorb it without violating policies.

Benefit versus risk

The engine ranks candidate actions by expected benefit in utilization or performance against the risk of contention if the action is executed.

Action approval workflow

In Manual mode, pending actions are listed for review; approving one causes the platform to execute it on the target using the automation credential.

Provision action

A provision action adds capacity, such as a host or node, when existing supply cannot satisfy demand without violating constraints after other actions.

Action not executing

When actions do not run, check in order the applicable action mode policy, approval status, credential permissions, and target connectivity.

Missing entities

Entities absent from the supply chain usually indicate a target discovery failure, missing stitching, or a target that was removed from the platform.

Audit and action history

History records which actions were proposed, approved, and executed, by whom and with what result, and is the primary trace for automation questions.

Time skew impact

Clock drift between the platform and targets can cause authentication and discovery failures, so NTP synchronization is a standard troubleshooting check.

REST API usage

The REST API exposes entities, actions, and policies for querying and allows actions to be approved or executed programmatically by external automation.

Webhook notifications

Webhooks push event and action notifications to external systems such as ticketing or chat tools, enabling workflow outside the platform UI.

APM integration

Application performance monitoring integrations supply application-level data so workload performance can be correlated with the underlying infrastructure entities.

Kubernetes operator

An operator-based deployment lets the platform run natively on Kubernetes or OpenShift and manage container clusters as targets.

Upgrade order

For multi-node environments, back up the database first, upgrade services and analytics in supported order, then re-verify targets and policies.

Backup and restore

Backups include the database and configuration; a restore must target the same platform version before any further upgrade is applied.

Discovery interval

Discovery runs on a target-specific schedule and can be triggered manually; the first run populates the supply chain and later runs refresh it.