Telegraf

Plugin-driven agent for collecting, processing, aggregating, and writing metrics, logs, and events from infrastructure, applications, and devices. Connects any data source to any destination via 400+ community-tested input, processor, and output plugins. Written in Go, compiles to a single binary with no external dependencies and minimal memory footprint.

Telegraf is the collector component of the TIG stack (Telegraf, InfluxDB, Grafana) and is maintained by InfluxData. It differentiates from pull-based scrapers like Prometheus with a push model suited for short-lived jobs and edge deployments. Telegraf handles buffering and backpressure to prevent data loss, supports DaemonSet and sidecar patterns in Kubernetes, and runs reliably on resource-constrained IoT and edge devices.

Key capabilities:

• 400+ input plugins for databases, clouds, containers, IoT protocols (MQTT, Modbus, OPC-UA), and system metrics
• Processor and aggregator plugins for filtering, enriching, and transforming data in flight
• Output plugins for InfluxDB, Prometheus, Kafka, Grafana, cloud monitoring, and 100+ destinations
• Single static binary with zero external dependencies, deployable at edge or in containers
• Built-in service discovery for Kubernetes, Consul, and static targets

Engineering teams use Telegraf for infrastructure and application monitoring, IoT and industrial telemetry, Kubernetes and Docker metrics, and DevOps observability pipelines. It pairs with InfluxDB and Grafana for full-stack time series workflows and integrates with Prometheus exporters, cloud provider APIs, and custom scripts via exec plugins.