Hybrid infrastructure monitoring

Know when an endpoint stops answering, before your customers do

CloudMonitor checks your URLs from outside, confirms a failure with a second probe before it counts as down, and listens for the cron jobs that stopped reporting. It sits alongside the health signals from the platforms in your environment, or your clients'.

What goes wrong

A service can look healthy from the inside and be unreachable from the outside: an expired certificate, a DNS change, a firewall rule, a backend that answers the platform's health probe but not a real request.

Scheduled jobs fail the other way. A backup or sync that stops running raises no error, because nothing is running to raise one.

Single-probe monitors cry wolf. One dropped packet pages someone at night, and after a few of those the alerts get muted.

How CloudMonitor catches it

Coverage by platform

Every platform CloudMonitor connects to, in the same order on every page, with what it checks for this problem today. Where nothing is checked yet, the table says so.

PlatformWhat CloudMonitor checks for this problemStatus
Microsoft Azure
Traffic Manager: Profile and endpoint health: every endpoint down is critical
Front Door & CDN: Custom-domain certificate expiry and validation, disabled endpoints and origins, WAF mode
Azure Monitor: Alerts still firing
Live
Microsoft 365 & Entra ID
Service Health & Message Center: Incidents and retirements
Live
Amazon Web Services
Route 53: DNS posture, domain expiry, health-check status
CloudWatch: Alarms in ALARM, failed status checks
AWS Health: Incidents and scheduled changes
Beta
Google Cloud
Load balancers: Backend health, external load balancers without Cloud Armor
Cloud Monitoring: Open alerts
Service Health: Incidents
Live
Kubernetes
AKS: Version support, node-pool health, upgrades
Amazon EKS: Version support, node groups, control plane
Google GKE: Version support, node pools
Live
Containers & registries
Docker Engine hosts: Privileged containers, restart loops, exposed ports, disk
Live
On-prem servers & Exchange
Internal TLS endpoints: Certificate expiry on any host and port, no credential
Linux servers (SSH): Disk, pending reboot, patch age, certificate files
Windows servers (WinRM): Disk, pending reboot, patch age, certificate store
Beta
MuleSoft Anypoint
CloudHub runtime: App status, unsupported runtime versions
Runtime Manager alerts: Alerts still firing
Live
Domains, DNS & email
TLS certificates: Expiry, TLS 1.0/1.1, weak ciphers
Live
Uptime & heartbeats
HTTP(S) endpoints: Reachability and TLS handshake every 5 minutes on paid plans
Heartbeats: Late or missing cron and job pings
Live

How deep each platform goes. Live means scheduled for every connected account; Beta means scheduled and verified against emulated APIs only. Azure is the most complete and is run against real accounts every day. The AWS and Google Cloud scanners have been verified against emulated APIs, not yet against live customer accounts. Kubernetes, MuleSoft and the on-prem collector are narrower still.

Connect your platforms

Access is read-only and least-privilege, and there are no agents to install on your cloud accounts.

Full setup guide: connectors.

Frequently asked questions

How often are endpoints checked?

Every 5 minutes on paid plans, and every 15 minutes during the free trial. 10-minute checks are available too.

How does it avoid false alarms?

A failure is confirmed by a second probe after a short pause before the endpoint is marked down, so a single dropped request does not page anyone.

Can it check things that are not websites?

Heartbeats cover scheduled jobs: your job pings a URL when it finishes and you are alerted if a ping is late or missing. Endpoint checks cover HTTP and HTTPS URLs.

How long is check history kept?

90 days.

Does it tell me when the certificate on an endpoint is about to expire?

Yes. Each check reads the certificate that was served and reports days left, with the same step-up alerts as every other expiry in CloudMonitor.

Is the platform health coverage the same everywhere?

No. The table lists what each platform contributes. The endpoint and heartbeat checks themselves work against any URL or job, wherever it runs.

Related

Try it on your environment, or your clients'

14-day free trial. No credit card needed to start, and it never turns into a paid plan on its own.