CloudStats.me
Start free
Log inStart free

Server availability alerts

A server availability alert watches whether the CloudStats Linux agent is still contacting CloudStats.

Updated Nov 8, 2026

On this page

Summary#

A server availability alert watches whether the CloudStats Linux agent is still contacting CloudStats. It is not a public ping, and it is not an alert that a metric went silent.

After the agent has been out of contact for the rule’s duration, CloudStats sends one server-down notice. When contact resumes, CloudStats sends one recovered notice. A server can appear offline in the CloudStats app sooner than an availability alert, because the app marks a server online when the agent has contacted CloudStats within the last 60 seconds.

Prerequisites#

You do not need a public ICMP ping target or an extra probe on the host.

Create a server availability alert#

In the CloudStats app, create an availability alert rule with these settings:

  1. Name — a label for the rule.
  2. Type — availability (not a metric threshold).
  3. Duration — how many seconds the agent must stay out of contact before CloudStats sends a server-down notice. Use a whole number of 0 or greater. 0 means CloudStats can alert as soon as the agent is treated as out of contact.
  4. Severity — Info, Warning, or Critical. Severity labels the rule and appears in notifications. It does not change when the rule fires.
  5. Servers — all servers, or selected servers.
  6. Notification channels — optional. If you attach no channels, the rule can still evaluate, but CloudStats does not send notifications.

You can later edit or delete the rule in the CloudStats app. You can also turn a rule off or on if that control is shown.

Expected behavior#

  • CloudStats decides availability from the agent’s last authenticated contact, not from ping, not from open ports, and not from whether a particular chart has recent points.
  • If the agent resumes contact before the duration elapses, CloudStats does not send a server-down notice and does not send a recovered notice.
  • When the agent has been out of contact for at least the duration, CloudStats sends one server-down notice. It does not keep resending that notice while the server stays down.
  • When contact resumes after a server-down notice, CloudStats sends one recovered notice.
  • The CloudStats app can show the server as offline after 60 seconds without contact, even if the availability rule’s duration is longer. Offline in the app is not the same moment as an availability alert.
  • An online status means recent agent contact. It does not guarantee that the latest chart point was stored. See Service, status, and logs.
  • Notifications use server.down and server.up events. See Notification channels.

Troubleshooting and limitations#

What you seeWhat it means
Server is offline in the app, but no server-down notice yetExpected when the offline window (60 seconds) is shorter than the rule’s duration. Wait until the duration has elapsed, or check that the rule has notification channels.
Agent is stopped, firewalled, or cannot reach CloudStatsThe availability rule watches contact with CloudStats. Check that cloudstats-agent is active and the host can reach https://api.cloudstats.me. See Troubleshooting.
No notice, and the agent came back before the duration elapsedExpected. CloudStats does not send a down or recovered notice in that case.
Charts look empty while the server is still onlineOnline status and chart storage are different signals. See Troubleshooting and Metrics history and chart ranges.

Limitations of the current product:

  • Availability alerts do not ping the host, check ports, or probe a website or TLS certificate.
  • They do not mean “this metric stopped reporting.” Use a metric alert rule to compare a host metric to a threshold.
  • If a server-down notice was sent but you did not receive it, see Troubleshoot alert and notification delivery.