Installation, Deployment, and Upgrade of Instana Agents and Hosts
This domain tests how Instana agents and host monitoring are installed, deployed, and upgraded across host, container, and Kubernetes environments. Expect questions on agent keys, endpoints, outbound connectivity, agent placement, upgrade strategy, and verification.
Deployment choices
A host agent is installed on each monitored host; it discovers processes and starts sensors. Containerized agents are used where the host cannot take a package, and Kubernetes clusters normally use the Instana Agent Operator, which creates a DaemonSet so every node gets an agent. The agent needs an agent key and a tenant-specific endpoint. Communication is outbound over TLS, normally port 443; proxies must be configured explicitly. Use zones and tags to separate prod, staging, and dev.
Upgrade discipline
Upgrades come through the OS package manager, the container image tag, or the operator custom resource. Follow the compatibility matrix, pin image versions in controlled environments, and upgrade the operator and CRD before changing its managed agent image. Keep a rollback image and verify the agent version in the UI after rollout.
Worked example
A 30-node Kubernetes cluster has no host data. Install the operator, create an InstanaAgent custom resource with the agent key in a Secret, set endpoint, zone, and tags, then confirm one DaemonSet pod per node and agent status Connected. If pods run but hosts do not appear, check outbound 443 and DNS before reinstalling.
Traps
Do not assume inbound firewall rules are needed; the agent initiates connections. Do not reuse one key without zones and tags, and do not upgrade the operator without checking CRD changes.