IBM Instana Observability Administrator - Professional flashcards

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

Agent download key

A tenant-specific string that authorizes downloading Instana agent packages from the repository. It appears in the install script and in the agent configuration file, and differs from the agent key that identifies the tenant at runtime.

Agent key purpose

The agent key identifies which Instana tenant an agent reports to. It is embedded in agent configuration, must be distributed to every host or cluster, and must be rotated and redistributed if compromised.

Kubernetes agent DaemonSet

The Instana Helm chart or operator deploys the agent as a DaemonSet, running one agent pod per node. It reads the agent key from a Secret and needs host mounts for container runtime access.

Agent upgrade rule

An agent must be at or above the minimum version supported by the backend release. Agents below the minimum are rejected or flagged, so upgrade agents before or alongside backend upgrades.

Dynamic Focus Query syntax

A boolean tag expression using AND, OR, NOT and parentheses, with equality and wildcard matching, for example entity.type:docker AND docker.name:web*. It scopes infrastructure views, dashboards, and alerts.

entity.type special tag

A built-in Dynamic Focus Query tag identifying the entity kind, such as host, docker, or process. It is commonly combined with other tags to scope queries to a class of infrastructure.

Remote agent configuration

Configuration pushed from the Instana UI that overrides supported local agent settings. It lets administrators change sensor and tag settings across many agents without editing each configuration file.

Sensor enablement flag

Each plugin section in agent configuration has an enabled true or false setting. Disabling a sensor stops that data collection while the agent continues running and reporting other metrics.

Host tags for scoping

Tags defined in agent configuration become entity tags used by Dynamic Focus Queries, alert scopes, and application perspective membership. Consistent tagging is required for reliable grouping and filtering.

Container runtime socket access

Docker and containerd monitoring requires the agent to read the runtime socket and host cgroup filesystem. Without socket access, container metrics are missing even though host metrics continue.

Agent heartbeat and offline hosts

A host appears offline when no agent heartbeat reaches the backend within the timeout. Check connectivity, proxy settings, and clock skew before assuming a backend fault.

Custom metrics endpoint

The local agent endpoint on port 42699 accepts custom metrics and events from applications. Submitted data is attached to the host entity and appears in infrastructure views.

Application perspective scope

A perspective is defined by a scope over services and endpoints plus boundary and alerting settings. A service appears in the perspective only when its tags match the scope.

Trace context propagation

Distributed tracing relies on standard headers such as W3C traceparent and tracestate. If a service fails to propagate them, the trace splits into separate unrelated traces.

Java tracing sensor

The JVM tracing sensor instruments Java applications automatically, typically by attaching an agent or using a javaagent option. It must load successfully for services and endpoints to appear.

Trace sampling rate

A per-sensor setting controlling what fraction of traces is sent to the backend. Lower rates reduce volume and cost but hide detail for rare or low-traffic request paths.

Alerting configuration components

An alerting configuration combines a metric or event condition, a scope, a threshold with a time window, a severity, and notification channels. A mismatch in any part prevents firing.

Alert silencing

A temporary suppression of notifications for a defined scope, used during planned maintenance. Silencing hides notifications but does not delete or disable the underlying alert configuration.

Notification channel types

Instana supports email, Slack, PagerDuty, Opsgenie, ServiceNow, webhook, and generic HTTP channels. Each channel is configured once with credentials and reused across alert configurations.

Smart alert grouping

Related issues from a common cause are grouped so one incident produces one notification. Grouping considers entity relationship and time proximity, reducing notification storms during outages.

Team-based visibility scoping

Application perspectives and infrastructure scopes can be restricted to specific teams. Users then see only entities granted to their teams, enforcing separation between business units or environments.

SAML and OIDC SSO

Instana integrates with identity providers using SAML 2.0 or OIDC. The IdP asserts identity and group membership, and Instana maps those groups to teams and roles.

API token versus user login

API tokens authenticate automation and carry permissions of their creator or a service account. They are separate from interactive SSO logins and should be scoped and rotated regularly.

Least privilege roles

Administrative roles should be limited to a small group. Monitoring roles grant read access to scoped entities without configuration rights, reducing risk of accidental or malicious changes.

Retention per data type

Metric, trace, and log retention are configured separately. Longer retention increases storage cost and query time, so align each data type's window with its diagnostic purpose.

PII scrubbing in traces

Capture rules and sensor settings can exclude request payloads, headers, or query parameters from traces. This prevents sensitive personal data from being stored in the observability backend.

Agent TLS and proxies

Agent traffic to the backend uses TLS on port 443. Behind an inspecting proxy, the agent must trust the proxy's certificate authority or connections fail.

Data residency region

The tenant's backend region determines where telemetry is stored and processed. Confirm the region before onboarding workloads subject to residency or regulatory constraints.

Agent diagnostic bundle

A support archive the agent generates containing configuration and logs for troubleshooting. Collect it before opening a case and redact secrets such as keys and tokens.

Clock skew impact

Incorrect host or container time breaks TLS validation and misaligns trace timestamps. Verify NTP synchronization on hosts and containers when connections or traces behave unexpectedly.

Proxy environment variables

Agents behind a proxy need HTTP_PROXY and HTTPS_PROXY set, plus NO_PROXY for local endpoints. Missing or misconfigured values cause silent failures to reach the backend.

Agent CrashLoopBackOff causes

A containerized agent repeatedly restarting usually indicates an invalid agent key, an unreachable backend endpoint, or a missing runtime socket mount. Check pod logs and the key Secret.