Monitor your hosts and containers.

Track CPU, memory, disk and network across hosts, containers, Kubernetes workloads and datastores. Every resource carries the service it belongs to, so you can open a service and see what runs under it.

Start free Read the docs

AgentDatadog Agent works as-is

Also readsOTLP, Prometheus

Grouped byService, environment and team

Hosts · prod48 hosts · by cpu

env:prodgroup by serviceLive

hostservicecpumem

prod-web-07checkout-api42%72%

prod-web-03checkout-api61%70%

prod-db-01postgres58%84%

prod-worker-04payments-worker47%52%

prod-web-11checkout-api39%66%

Hostprod-web-07

servicecheckout-api

regioneu-central-1a

instancec6i.xlarge · 4 vcpu

uptime41 days

Loadnominal

cpu42%

memory72%

disk38%

net out283 Mb/s

Processesby cpu

node · checkout-api39%

moneat-agent1%

envoy1%

Check the infrastructure under a failing service

Infrastructure answers a narrow question during an incident: is the platform the cause, or is it affected too. Here is one service, from its hosts to the metric that confirms the fix.

INC-402 · Checkout latency above SLO

Resolved

14:02:11 → 14:19:44 · 17m 33s

14:03:50

Hosts under the service, already filtered

Six hosts and their containers, scoped to the service named in the alert.

Infrastructure

14:05:30

Throttling found, saturation ruled out

CPU headroom is fine, the container is being throttled, and the database is at its connection limit.

Containers

14:12:10

Pool ceiling raised, metrics follow

Connections in use fall away from the limit within a minute of the change rolling out.

Metrics

Track the resources your services run on

Hosts, containers, Kubernetes workloads and datastores, collected by an agent you may already be running.

Host metrics

CPU, memory, disk, network and load, per host, with the processes that are consuming them.

hosts

Containers

Per-container CPU, memory, restart count and CPU throttling, which a CPU usage graph alone does not show.

containers

Kubernetes

Clusters, nodes, workloads and pods, with events and evictions on the same timeline as the metrics.

kubernetes

Datastores

PostgreSQL, MySQL and Redis: connections, replication lag, slow queries and cache hit rate.

integrations

Scoped by service

Resources carry the service, environment and team they belong to, so you can narrow to one of them.

facets

One agent, many signals

A Datadog Agent reports metrics, host signals and traces here without changes to its configuration.

agent

Resource catalogue

One searchable inventory of everything you run, with its owner and what depends on it.

catalogue

Open a host and see three things in order.

A host page is ordered around the questions you ask when you open one: what is running on it, what changed, and whether it explains what you are investigating.

What is on it. Containers, processes and the services they belong to, ranked by consumption.
What changed. Deploys, restarts, evictions and config changes on the metric timeline.
What it talks to. Dependencies drawn from the traces the host actually served.
Open a request. From a busy host to the traces that were running on it.

See traces

ip-10-0-4-2214:02 → 14:20

Instance: c6i.2xlarge · eu-west-1a
CPU: 41% (headroom fine)
Memory: 11.2 / 16 GiB
Containers: 4 (1 throttled)
Restarts, 1h: 0
Last deploy: 13:58 · 2f91c
Traces, 1h: 1 284 · 24.1% errors

Related features

Traces & APMThe requests that were running on the host you are looking at.Alert routesAlert on saturation and throttling without firing on every spike.DashboardsChart these metrics alongside data from your own databases.

Start monitoring your infrastructure.

Point an agent at Moneat, or send OTLP and Prometheus metrics, and see your hosts and containers grouped by service. Start on the free plan with no card.

Start free See pricing