Sematext Agent

To investigate a cluster problem, a team may need to correlate application symptoms with host and container measurements. Sematext Agent gathers that infrastructure telemetry and sends it to the configured Sematext service, where it feeds monitoring and related views. On Kubernetes it runs across worker nodes, so collection scope and node permissions matter. Decide which data should leave the cluster, which regional endpoint receives it and how the resulting telemetry supports the alerts your team actually uses.

Current guidance

This entry identifies the general Sematext Agent rather than the deprecated Docker-only agent. Current Sematext documentation installs a DaemonSet from its own Helm repository and supports an existing Kubernetes Secret for the infrastructure token. Prefer the vendor installation guide over the repository README’s leftover stable-chart example, and pin the selected chart and image rather than inheriting a moving latest tag.

The documented agent runs as a privileged container. Review that requirement against node isolation, admission policy and the sensitivity of logs and process metadata collected. Select the correct US or EU destination and confirm which telemetry types are enabled; infrastructure monitoring and log shipping can have different data and billing consequences.

For migration, preserve filters, labels, application mappings, regional endpoints and alert dependencies. Compare a bounded set of nodes in the target account, checking for duplicate ingestion and missing containers before removing the old collector. Keep tokens out of command history and generated release files where possible. The documentation’s historical minimum Kubernetes version is not a certification of every newer cluster or hardened node configuration.