Infrastructure
Infrastructure is the fleet-level view of everything running under your applications. The overview at /observability/infrastructure reports health across Kubernetes, data platforms, hosts, and cloud scrapes, and links into per-asset surfaces for Hosts, Databases, Data stores, Messaging, Network devices, and Cloud.
What it is
Application telemetry answers "is the service healthy?"; infrastructure telemetry answers "is the machine, database, or network underneath it healthy?" Owlpane keeps both in one workspace, so a latency spike in a service can be traced down to a saturated node, a slow query, or a dropped network flow without switching tools.
The overview rolls up the fleet; each sub-surface goes one level deeper — aggregated and then per instance.
Key capabilities
- Fleet overview — Kubernetes, databases, data stores, messaging, hosts, and cloud scrapes in one health snapshot for the selected range.
- Hosts — node and VM metrics, both aggregated and per host.
- Databases — PostgreSQL, MySQL, and MongoDB metrics with per-instance detail.
- Data stores — Redis and in-memory cache metrics.
- Messaging — Kafka and RabbitMQ views for broker and topic health.
- Network devices — an interface inventory, device topology, and flow views, so you can move from a device to the traffic crossing it.
- Cloud — scraped cloud provider resources, with provider-specific pages for AWS, GCP, and Azure.
- Operations — curated alert rules and maintenance windows to keep infrastructure noise down.
Where it lives
| Surface | Console route | What it does |
|---|---|---|
| Overview | /observability/infrastructure | Fleet health across Kubernetes, data platforms, hosts, and cloud scrapes. |
| Hosts | /observability/infrastructure/hosts (/hosts/:name) | Node and VM metrics, aggregated and per host. |
| Databases | /observability/infrastructure/databases (/databases/:instanceId) | PostgreSQL, MySQL, and MongoDB metrics and instance detail. |
| Data stores | /observability/infrastructure/data-stores | Redis and in-memory cache metrics. |
| Messaging | /observability/infrastructure/messaging | Kafka and RabbitMQ integration views. |
| Network devices | /observability/infrastructure/network-devices (/topology, /flows) | Interface inventory, device topology, and flow views. |
| Cloud | /observability/infrastructure/cloud (/aws, /gcp, /azure) | Scraped cloud provider resources, per provider. |
| Maintenance | /observability/infrastructure/maintenance | Maintenance windows that suppress infrastructure noise. |
| Kubernetes 360 | /observability/infrastructure/kubernetes | Cluster hub for logs, metrics, traces, and incidents. |
How data gets in
Infrastructure telemetry comes from collectors and integrations rather than from application code: host metrics collectors, cloud scrapes, database exporters, and network flow sources, all forwarding OTLP or remote-write to ingest.owlpane.com. Kubernetes telemetry is covered separately in Kubernetes.
- Integration library — host, database, messaging, and cloud packs.
- SDKs — Runtime telemetry boards for application runtimes.
- Kubernetes Helm agent — node, pod, and host metrics from a cluster.
- Metric ingestion — how the collected series are stored and queried.
- OTLP ingest reference — forwarding endpoints.
Concepts
| Term | Meaning |
|---|---|
| Host | A node or VM reporting system metrics. |
| Instance | One database or data-store deployment. |
| Scrape | A pull of metrics from a cloud provider or an exporter. |
| Flow | Network traffic observed between endpoints or devices. |
| Maintenance window | A period during which infrastructure alerting is suppressed. |
How the layers relate
Application telemetry and infrastructure telemetry are two views of the same request. A latency spike in Traces can be followed down to a host, a database instance, or a network flow on this surface, and back up again. The overview deliberately rolls up every category so a single page can tell you which layer is misbehaving before you drill in.
Related
- Kubernetes — the cluster-specific surfaces.
- Metrics — the underlying series and how to query them.
- Dashboards — host, database, and cloud packs from the library.
- Alerts and on-call — infrastructure rules and maintenance.
- Carbon — energy and emissions attributed to the same fleet.
Last updated