Bindplane Makes Its AI Agent for OpenTelemetry Pipelines Generally Available
SALT LAKE CITY -- Bindplane used KubeCon + CloudNativeCon North America 2026 to make Bindplane Agent, the AI assistant built into its OpenTelemetry-based telemetry pipeline platform, generally available after an early access period. The company also introduced a new Overview page, first-class support for gateway collectors and Edge Fleets, and Full-Pipeline Blueprints in its configuration wizard.
Bindplane Agent lets users describe a desired outcome in plain language, such as sending Kubernetes logs to a specific observability destination, and then adds the sources, destinations and processors to the pipeline graph as users watch. Nothing deploys until a user rolls the change out. The agent also works in reverse, walking an existing pipeline to explain which sources send logs and where they go, or describing processor bundles that someone else built. It is available to every organization with Pipeline Intelligence enabled, and an org admin turns it on.
New Overview Page Shows Data Flow, Topology and Fleet Health
The new Overview page has three tabs. Data Flow shows configurations, gateways and destinations with each connection sized by throughput. Topology presents the same deployment as a graph, with destinations that have stopped receiving data shown greyed out. Summary covers collector health, rollouts, version spread and a throughput chart comparing data in against data out over 24 hours, 30 days or 90 days.
Gateways and Edge Fleets Become First-Class Resources
Bindplane now assigns each OpenTelemetry Collector a role, either agent or gateway. Gateways get their own navigation entry, table and install flow, with Kubernetes recommended as the deployment target and the option to have the OpenTelemetry Operator deploy them. A gateway configuration can only be assigned to a gateway, so an edge configuration cannot land on the aggregation tier by mistake. Edge Fleets group collectors on VMs, workstations and Kubernetes nodes, and configuration assignment follows a fixed order of gateway configuration, then fleet, then label.
Full-Pipeline Blueprints now appear as the first option when creating a configuration. Each blueprint bundles a source, processors, routing, batching and a destination tuned for that source, with parsing, filtering, masking and sampling already in place. The largest includes 32 processors, which users can review and add in a single step instead of building each one by hand.
