Every host andcontainer onone page.

One agent per host takes its pipeline from the portal. CPU, memory, disk and network arrive without an exporter per service, and the integration catalogue covers the databases, queues and proxies underneath.

insight / infrastructure monitoringLIVE
HOSTCPUMEMVAL
ip-10-0-1-42
184 MB/s
payments-api-2b1d
41 MB/s
postgres-primary
210 MB/s
redis-cache-0
18 MB/s
kafka-broker-1
96 MB/s

Read from the kernel on each host. No exporter per service to install or keep current.

Collected fromTrademarks →
What you get

Included on every plan, including the free tier.

01

Host and container resource metrics, collected without a per-service exporter to install and version.

02

More than ninety integrations covering PostgreSQL, MySQL, Redis, Kafka, nginx, RabbitMQ, Elasticsearch and the rest of the estate.

03

Alert rules and dashboards on the same data, so a threshold and the graph behind it never drift apart.

04

The same tenant your logs and traces land in, so a saturated host and the slow request on it sit side by side.

How it gets there

Three steps, then it is reporting.

01Add one opamp extension block to a collector on the host.
02Assign the host pipeline from Fleet Management.
03Hosts appear in Insights with dashboards attached.
What you can ask

What that looks like as a query.

PROMQL

topk(5, 100 - avg by (instance) (rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100)

The five busiest hosts right now.

PROMQL

node_filesystem_avail_bytes / node_filesystem_size_bytes < 0.1

Filesystems inside ten percent of full.

PROMQL

rate(pg_stat_database_xact_rollback[5m]) > 0

Databases rolling transactions back.

Questions

What platform teams ask before the rollout.

Something not here? Book thirty minutes with an engineer who works on the platform. Book a call →

Do we have to replace Prometheus?
No. Point its remote_write at us and your existing scrape jobs keep running. Recording rules, alert rules and PromQL are unchanged, and Grafana Alloy forwards the same way.
How many agents per host?
One. It is a standard OpenTelemetry Collector with a single opamp block; every receiver, processor and exporter after that arrives from the portal, versioned, with secrets resolved on the agent.
What about the databases and queues?
Each integration is a receiver plus a dashboard. You assign it to the agents that should run it rather than deploying another exporter per service.
Will this cost more than what we run now?
Cardinality decides that, not host count. Labels carrying unbounded values grow series faster than traffic does, and the agent can drop them before they are billed.

Start on the free tier.

20K active series, 50 GB of logs and 50 GB of traces a month, 90-day retention, and the whole of Insights.