The 2 AM pages
Credential expiry
A secret, API key or certificate reaches its expiry date and every API call that depends on it fails — simultaneously.
Idle VMs on autopilot
Every cloud makes it easy to spin things up. None of them remind you to spin them down. Forgotten instances bill $300–$800/mo each, across every account you own.
Silent endpoint failures
Your status page says green. Customers start reporting errors. Somewhere, an HTTPS endpoint went down two hours ago.
What it watches
Everything that quietly breaks across your cloud stack
Secret & certificate expiry
Tracks expiry across Azure Key Vault, AWS Secrets Manager and ACM, and checks rotation on GCP Secret Manager. Alerts step up as the date approaches — 90, 60, 30, 14, 7 and 1 day out, and on the day itself.
Secrets · Certs · Keys
Compute cost analysis
Identifies idle and underutilized Azure VMs, AWS EC2 and GCP Compute Engine instances from CPU and memory telemetry over 7-day windows. Reports estimated monthly waste in dollars, not percentages.
VMs · EC2 · Compute Engine
Endpoint monitoring
HTTP/HTTPS uptime checks every 5 minutes on paid plans (every 15 during a trial), with response-time tracking and TLS certificate expiry alerts. A failure counts as down only when a second probe about 20 seconds later confirms it, and every check is kept for 90 days with its status, response time and error, so there is a record of when an outage started and ended.
HTTP · HTTPS · SSL
Storage hygiene
Flags unattached managed disks, aging snapshots, and publicly readable Azure Blob containers, S3 buckets and GCS buckets — with estimated monthly cost and a link straight to the resource.
Blobs · S3 · GCS · Disks
Microsoft 365 & Entra ID
Users with no MFA, stale and guest accounts, privileged roles, Conditional Access gaps, app-registration secrets and SAML signing certificates near expiry, unused licenses, Intune Apple tokens, and Secure Score with its trend. The directory connects on its own — no Azure subscription needed.
Entra · Intune · Secure Score
Domains, DNS & email
Domain registration expiry, TLS expiry and weak TLS, SPF, DKIM, DMARC and MTA-STS, DNS drift, DNSSEC, CAA and dangling DNS records. Certificate Transparency watch flags certificates from a CA you never approved, and lookalike domains registered against yours.
Domains · DNS · Email auth · CT
Cost & governance
Reservation and Savings Plan expiry, cost anomalies and budget burn, quota headroom, operating systems past vendor support, missing tags, policy and resource locks, plus Defender and Advisor recommendations and backup coverage.
Costs · Quotas · EOL · Policy
GitHub security & CI/CD (Beta)
Default-branch protection, open Dependabot, code-scanning and secret-scanning alerts, deploy keys and token expiry, with a CI/CD board for workflows that are failing, stuck or flaky. A read-only token; it never re-runs a workflow or reads your code.
Repos · Actions · Alerts
Beyond the scan
Findings you can act on, wherever you work
Diagnostics
Related signals are fused into one likely cause per problem, so an expired secret, the failing app that depends on it and the endpoint that went down read as one story rather than three alerts.
Alerts where you already work
Email, Slack, Microsoft Teams, Discord, PagerDuty or your own webhook, and tickets in ConnectWise, Autotask, HaloPSA or NinjaOne — all included. Choose, per destination and alert type, whether alerts arrive at once or in an hourly, daily or weekly digest, and whether you hear when they resolve.
One environment or fifty
Watch your own environment, or every client's. For MSPs: each client in its own isolated tenant under your MSP account, one cross-client worklist of what is critical or expiring, one-time onboarding links for clients, and a read-only portal for each of them.
Open by default
A read-only REST API on every plan, single sign-on with your own identity provider over OpenID Connect, unlimited users, an activity log, and a full export of your data whenever you want it.
Platform coverage
CloudMonitor has no single vantage point to watch from, because there isn't one. It walks every platform in turn, checking each resource on the way past and coming back around on a schedule: endpoints every 5 minutes on paid plans, exposure across the cloud accounts every hour, and expiry, rotation and cost once a night — the 2 AM round, made so you can sleep through it. The map below is the route.
Coverage is uneven, and it would be misleading to present it otherwise. Azure is where the depth is. AWS and Google Cloud are broad but newer, and the rest are real and still growing. Azure is the platform CloudMonitor has been run against for real, every day. The AWS and GCP scanners have so far been verified against local emulators of those clouds rather than live accounts, so expect the rough edges to show up there first.
How it connects
No agents to install and no policy changes. CloudMonitor connects with least-privilege read-only access — on Azure a service principal or a secretless federated connection (Beta), on AWS a key for a dedicated read-only IAM user, on GCP a read-only service-account key, each stored encrypted — and never needs write permission on anything it watches. Cross-account roles for AWS and workload identity federation for GCP are planned, not available yet. Each customer's data sits in its own isolated tenant.
-
1
Grant read-only access
Per-platform onboarding guides walk through the exact role and scope, so you can review it before you grant anything.
-
2
Scanners walk the accounts, pass after pass
Every pass re-enumerates what is actually there, so findings describe the account as it stands rather than as it looked on the day you connected it. They are ranked by severity and estimated cost impact, not dumped as an undifferentiated list.
-
3
Alerts where your team already looks
An endpoint that stopped answering is reported within minutes of the check that caught it; expiry thresholds and new findings such as a public bucket are reported after the nightly scan that found them. Alerts go out by email, and can also post to Slack, Microsoft Teams, Discord, PagerDuty or a webhook you set up, or open a ticket in ConnectWise, Autotask, HaloPSA or NinjaOne; you choose which kinds of alert you receive.
Pricing
From $29/month, priced in bands by how much you monitor. Size is counted in units: a core resource — a VM, database, cluster, app, load balancer or connected SaaS platform — is 1 unit; a supporting one — a secret, certificate, endpoint, container, bucket, disk, DNS zone, NSG or public IP — is 0.2 of a unit. Every account starts with a 14-day free trial, no credit card needed.
Every band includes every scanner and alert channel, unlimited users, and read-only API access. AI Insights is included from Team up. Three optional add-ons are priced separately:
“The best monitoring tool is the one that catches the problem before your users do. CloudMonitor was built by someone who has been paged at 2 AM over an expired certificate.”
Cloud architect · 20+ years in IT, 15+ in the cloud
Your infrastructure deserves a second pair of eyes.
CloudMonitor is open. Connect your first cloud account and see what it finds; the first 14 days are free, and no credit card is needed to start.
Start your free trial