Pricing that scales with your stack

A service is one monitored node — a single Elasticsearch, Redis or Logstash instance is one service. Disk and host/system metrics are always included free on every VM and never count against your limit, so even the Trial plan gets full disk-fill forecasting plus one real service. A cluster counts one service per node — because we track every node independently, an 8-node Elasticsearch cluster is 8 services. The convenience is that one agent still collects the whole cluster, so there is no per-node install or config to maintain; the per-node count only affects how it tallies against your plan.

Each tier gets a different annual discount: Starter is 1 month free (12 paid months for the price of 11), Pro is 45 days free, and Premium is 2 months free. The discount is applied as a lower annual price; if you switch from monthly to annual mid-cycle, the remaining months are credited.

Nothing breaks and nothing is dropped mid-incident. Limits are soft: you can keep adding services and we simply flag in-app that you are over. If you stay over for an extended period without upgrading, over-limit services may be paused — but you are always warned first, and pausing never deletes history. We would rather you see the alert than hit a wall during an outage.

Both. Every paid plan includes a 14-day free trial with no card required up front. Beyond that, the free Trial plan keeps working for a single service indefinitely — enough to forecast disk fill on one VM or watch one Redis instance for free, forever.

Yes, any time, self-service. Upgrades take effect immediately and are prorated; downgrades apply at the end of the billing period so you keep what you paid for. Cancel and your data is retained for 30 days in case you come back.

Usually under a minute. You run one curl command, the agent auto-detects your services, and the first metrics arrive within a collection interval (2 min on Trial, down to 10 s on Premium and Enterprise). Forecasting insights — like disk fill dates — sharpen over the first few hours as we collect enough history to model the trend.

No. There are no Prometheus exporters or sidecars to wire up. A single agent — Augur, built on open-source Telegraf — collects every supported service on the host. Add a new service later and the same agent picks it up — there is no second install.

Thresholds tell you a number crossed a line, after it crossed. Foreseer models the trend and tells you when a line will be crossed, with hours of lead time, plus the likely cause and the command to fix it. A disk at 80% is not an alert; a disk that will hit 100% by 2am tonight is.

No. The config we serve has placeholders, not passwords — you type the credentials for Elasticsearch, Redis or Logstash directly into the agent config on your own VM. They live only on that host, are never transmitted to or persisted on our servers, and are carried forward automatically across agent updates. We only ever receive the resulting metrics.

Numeric metrics and lightweight metadata (service versions, cluster names, host and cloud labels) — never log contents, document bodies, or key values from your datastores. Metrics are sent over TLS, authenticated with a per-host bearer token you can revoke at any time.

Metrics flow into a VictoriaMetrics-compatible store through the ingest API, so the time-series layer can run in your own environment. Fully self-hosted deployments are available on Enterprise — talk to us about your requirements.

Three roles: admins manage everything, project managers manage the services and alerts for the VMs they are granted, and viewers get read-only access. On Pro and above you can scope a manager to specific VMs, so a team only sees the hosts they own.

Email on every plan (including Trial). Slack and webhook destinations — PagerDuty, Opsgenie, or any custom endpoint — unlock on Pro. Connectors are reusable: create one destination and attach it to any service.