
Getting started
New to CloudMonitor? Start here.
What is CloudMonitor?+
CloudMonitor is a hybrid infrastructure monitoring SaaS built for cloud engineers and DevOps teams. It watches the environments you connect — Azure, Microsoft 365, AWS, GCP, Kubernetes, Docker and MuleSoft, plus your domains and endpoints (AWS is in Beta) — for problems that quietly cause outages — expiring certificates and secrets, idle VMs burning budget, endpoint failures, storage security issues, and more — and alerts you before they become incidents.
What platforms does CloudMonitor support?+
Microsoft Azure (26 services, including Key Vault, VMs, storage, databases, App Service, Container Apps, networking and cost), Microsoft 365 and Entra ID (users and MFA, admin roles, Conditional Access, app-registration secret and certificate expiry, licenses, Secure Score), Kubernetes (AKS, EKS and GKE), Docker (Docker Engine hosts and registries: Docker Hub, GHCR and self-hosted Registry v2 such as Harbor, Nexus, Artifactory, GitLab and Quay), Google Cloud (19 service groups), MuleSoft Anypoint Platform (CloudHub runtime, Runtime Manager alerts), and, with no credentials at all, your domains, DNS and email authentication, TLS certificates and endpoint uptime — all included on every plan. Amazon Web Services (21 service groups) is in Beta: verified against emulated AWS APIs; not yet validated on a live customer account. GitHub and Cloudflare are live in Beta; other SaaS platforms are coming soon. More platforms are added regularly —
contact us if you have a specific integration request.
How long does onboarding take?+
Connecting usually takes 5–10 minutes. Each platform has its own connection guide inside the dashboard, using least-privilege read-only access: for example a Service Principal with Reader access on Azure, a read-only IAM user on AWS, a service account with viewer roles on GCP. Endpoint checks begin within 5 minutes of adding a URL on a paid plan (15 during a trial). Exposure findings — public access, open network rules, cluster and service state — appear within the hour. Expiry and rotation findings appear after the next nightly scan, so expect your first full set the following day.
Do I need a credit card to start the free trial?+
No. Your 14-day free trial starts when you sign up and verify your email, so connect your first subscription or platform early to get the full 14 days. You only need to enter payment details if you decide to continue past the trial period.
Are AWS and GCP supported?+
Yes. GCP is live. AWS is in Beta: verified against emulated AWS APIs, not yet validated on a live customer account. Both are available on every plan. AWS coverage spans 27 service groups, including EC2, RDS, S3, IAM, Lambda, ACM, Secrets Manager, GuardDuty and Security Hub, load balancers, Route 53, CloudTrail, backup and cost. GCP coverage spans 25, including Compute Engine, Cloud SQL, Cloud Storage, BigQuery, IAM, Secret Manager, Cloud Run, Cloud DNS and Service Health. Each scanned resource adds to your unit count exactly as an Azure resource does — 1 unit for a core resource, 0.2 for a supporting one.
My service or platform isn't on the supported list — can you add it?+
Yes, we accept connector requests. Sign in and submit a
🔌 Connector Request from
Help → Support (
/index.php?view=support), describing the service and what you need monitored (resources, alert conditions, expected volume). It arrives as a support ticket and is read by the person who would build it. If approved, the connector is covered by the same unit bands as everything else (what it scans adds to your unit count like any other resource) and shipped through our normal release pipeline.
One limit: we don't take on connectors for services associated with illegal activity or with a clear malicious-use risk to our platform. If you don't yet have an account, you can reach us via the public
Contact page instead.
◫ Azure monitoring
How CloudMonitor connects to and scans your Azure environment.
What does CloudMonitor monitor in Azure?+
CloudMonitor scans 26 Azure services, including Key Vault (secret, certificate and key expiry, rotation policy), VMs (idle and underutilized, flagged by cost), Storage (public containers, unattached disks, aging snapshots), SQL, PostgreSQL, MySQL and Cosmos DB, App Service, Container Apps, container registries, network security groups, DNS, Defender, Monitor alerts, the Activity Log, backup and cost. Connect Microsoft 365 and Entra ID as well for user MFA, admin roles, Conditional Access, app-registration secret and certificate expiry, licenses and Secure Score. Endpoint uptime and SSL checks run alongside.
How does CloudMonitor connect to my Azure subscription?+
You provide a service principal — or, in Beta, add CloudMonitor's secretless application to your directory, with no secret to store — holding the Reader role on the subscription and Key Vault Reader on any vaults you want monitored. CloudMonitor never requests write permissions. An onboarding guide with exact Azure CLI and portal steps is included in the dashboard.
Does CloudMonitor read my secret values or key material?+
No. CloudMonitor reads metadata only — secret names, expiry dates, enabled/disabled status, and rotation attributes. It never reads the actual value of a secret, the content of a certificate, or any cryptographic key material. This is enforced at the API permission level.
How many Azure subscriptions can I connect?+
Unlimited on every plan. Connect as many subscriptions as you need — a subscription costs nothing in itself; the resources inside it add to your unit count (core resources 1 unit each, supporting ones 0.2).
Does CloudMonitor link to the resource in my cloud console?+
Yes. A finding links straight to its resource in the Azure Portal, the AWS Management Console or the Google Cloud console, so you can fix it without searching for it. A link appears only when we can build it from the finding itself; GovCloud and China regions get none, because their consoles live on different domains.
How are expiry thresholds configured?+
Findings are classified at 30 days (warning) and 7 days (critical) for Key Vault items and SSL certificates. Those levels are fixed. Rotation thresholds on the Alerts & notifications page add a NO EXPIRY or OVER POLICY label to rows in the Key Vault findings list, per resource type; the label does not change a finding’s severity or send an alert. Alert emails follow a fixed ladder — 90, 60, 30, 14, 7 and 1 days ahead, then on the day itself — and on the Alerts & notifications page you choose how far ahead the first notice starts.
⬡ MuleSoft / Anypoint
Monitoring CloudHub runtimes and Anypoint Platform alerts.
What does CloudMonitor monitor in MuleSoft?+
CloudMonitor connects to your Anypoint Platform and monitors CloudHub application runtime health (running, stopped, failed, deploying status for each app) and Anypoint alert findings (active alerts from Runtime Manager, ranked by severity). You get a unified view alongside your Azure and Kubernetes findings.
How do I connect my Anypoint Platform?+
In the dashboard, navigate to MuleSoft → Connections and click "+ Connect Anypoint Platform." You'll enter your Anypoint URL, Organization ID, Environment ID, and a Connected App client ID / client secret. CloudMonitor uses these to poll the Anypoint API — your credentials are encrypted at rest and never logged.
Which Anypoint environments are supported?+
Any Anypoint environment you have credentials for — Production, Sandbox, or custom-named environments. You can connect multiple environments from a single Anypoint organization and label them separately in the dashboard.
Does CloudMonitor support on-premises MuleSoft (Runtime Fabric / standalone)?+
Currently CloudMonitor monitors CloudHub-hosted applications via the Anypoint API. On-premises Runtime Fabric and standalone Mule runtimes are on the roadmap.
Contact us if this is a priority for your team.
⬢ Kubernetes
Cluster health monitoring across your K8s fleet.
What does CloudMonitor monitor in Kubernetes?+
For AKS, EKS and GKE, CloudMonitor reads the cloud provider's management API: cluster version and end-of-life, upgrade availability, node pools, and cluster configuration. With the in-cluster connector (Beta) it also reads the Kubernetes API directly: node readiness and pressure, unavailable workloads, crash-looping and image-pull-failing pods, restart counts, :latest images, missing memory limits, privileged and host-access workloads, namespaces without a NetworkPolicy, Ingress TLS coverage, cert-manager certificate expiry, cluster-admin bindings and wildcard roles, ResourceQuota headroom, Velero backup recency, and capacity and rightsizing (under-used nodes and node pools, and, with metrics-server, over-requested and idle workloads). Findings are ranked by severity alongside your other platform data.
How does CloudMonitor connect to a Kubernetes cluster?+
Two ways. AKS, EKS and GKE use the Azure, AWS or GCP connection you already have — nothing is installed in the cluster. Any other distribution uses the in-cluster connector (Beta): you apply a read-only ServiceAccount and ClusterRole (get and list only, no Secrets), and give CloudMonitor the API server URL, the token, and your CA certificate if you use a private CA. CloudMonitor never modifies cluster resources, schedules workloads, or changes configuration.
Can I connect multiple Kubernetes clusters?+
Yes. You can connect multiple clusters and label them by environment (Production, Staging, etc.) or region. All clusters appear in a single dashboard view.
Does CloudMonitor support on-premises Kubernetes or only managed services (AKS, EKS, GKE)?+
Managed Kubernetes — AKS, EKS and GKE — is scanned through the cloud provider's management API. Self-hosted and other distributions (OpenShift, k3s, Rancher, kubeadm) are supported through the in-cluster connector, currently in Beta, when the API server is reachable from the public internet over https. A cluster that is only reachable on a private network uses the CloudMonitor agent (Beta): a small outbound-only Deployment that runs inside the cluster and pushes findings to CloudMonitor — nothing connects in, and the only firewall requirement is outbound HTTPS to CloudMonitor. The agent image is private and is pulled from CloudMonitor with a per-cluster credential you control.
◑ Alerts & notifications
How CloudMonitor surfaces and delivers findings.
How do alerts work?+
When a scan detects a new finding — an expiring certificate, a failed CloudHub app, an idle VM — CloudMonitor creates a finding record and fires an alert through your configured channels. Alerts are deduplicated per item, so you won't receive repeat notifications for the same finding on every scan cycle, and you control how often a still-open alert may repeat. Anything that expires follows a fixed escalation ladder — 90, 60, 30, 14, 7 and 1 days ahead, then on the day itself — and each step is sent as soon as it is crossed, because it is new information. When the problem clears you get a recovery notice, so “fixed” never has to be inferred from an alert that simply stopped arriving; you can switch recovery notices off for any destination.
What alert channels are supported?+
Alerts are delivered by email, to Slack, Microsoft Teams, Discord or PagerDuty, or to a webhook URL of your own, and can open tickets in ConnectWise, Autotask, HaloPSA or NinjaOne — all included on every plan. For each destination and alert type you choose whether alerts arrive at once, in an hourly, daily or weekly digest, or not at all, and whether you are told when they resolve; PagerDuty and ticketing are always sent at once. Digest times follow the time zone you choose, UTC by default. There is no native Slack app; Slack is reached through an incoming webhook.
Can I customize alert thresholds and severity levels?+
Partly. The warning and critical levels are fixed: for Key Vault items and certificates, 30 days and 7 days before expiry. What you choose is when and where you hear about them: on the Alerts & notifications page you set how far ahead the first expiry notice starts and, for each destination and alert type, whether alerts arrive at once, in a digest or not at all. Rotation thresholds on the same page add a NO EXPIRY or OVER POLICY label to rows in the Key Vault findings list; they do not change a finding’s severity or send an alert.
How quickly does CloudMonitor detect a new issue?+
Endpoint uptime and certificate checks run every 5 minutes on a paid plan, and every 15 minutes during a trial. A failed check is repeated about 20 seconds later and only a second failure counts as down, so a single blip does not page you; once confirmed, the alert is sent in that same run rather than waiting for a separate delivery pass. Cloud scans run at three speeds, because the things they find move at different speeds. Findings your cloud provider has already raised — threat findings from Defender, GuardDuty, Security Hub or Security Command Center, high-risk changes in the Azure Activity Log, provider outages from Azure Service Health, AWS Health and Google Cloud Service Health, and incidents on a connected Statuspage — are read every 5 minutes on a paid plan (every 15 minutes during a trial). Anything else that can change without warning — storage or database public access, network rules open to the internet, container and cluster state, an app exposed without authentication — is re-checked every hour across Azure, AWS, Google Cloud and Kubernetes. Either way the alert is sent in that same run to every destination set to deliver at once; a destination set to a digest gets it in its next digest. Everything whose clock runs in days rather than minutes — certificate, secret, key and domain expiry, rotation age, end-of-life dates, utilization and cost — is scanned once a night, which is the right cadence for a warning ladder that steps at 90, 60, 30, 14, 7 and 1 days. So an endpoint problem, a provider-raised threat or an outage is normally picked up within 5 minutes on a paid plan (15 during a trial), an exposure within the hour, and an expiry on the next night’s scan. Apart from the 5-minute checks, which come with a paid plan, these cadences apply on every CloudMonitor plan. Some of these sources need a service or support plan from your cloud provider — see “Do any checks need a paid plan from my cloud provider?”
Can I silence an alert?+
Yes, one alert at a time. Under Open alerts on the Alerts & notifications page you can acknowledge an alert, which stops it repeating at its current step until it gets worse or an expiry reaches its next step; snooze it until a time you choose, up to 30 days, unless it gets worse, although the 1-day and expiry-day notices of anything that runs out are always sent; and assign it to the person handling it. Email, Slack, Teams, Discord, PagerDuty, webhooks and PSA tickets all honor the same choices, the daily digest lists what is acknowledged or snoozed and by whom, and all of it clears by itself when the problem is fixed. You can also switch each type of alert on or off, choose how far ahead expiry warnings begin, and set how often a still-open alert may repeat. Alerts are deduplicated per item, so an unchanged finding does not re-notify on every scan.
◎ Security & privacy
How CloudMonitor protects your data and your infrastructure.
What data does CloudMonitor store about my infrastructure?+
CloudMonitor stores metadata only — resource names, expiry dates, utilization metrics, endpoint response times, and finding records. It never stores secret values, key material, certificate private keys, or any application data from your workloads.
How is tenant isolation enforced?+
Every customer is assigned an isolated tenant ID, and all database queries, API responses, and UI views are scoped to your tenant, so no other customer can see your data. The operator's own staff roles can see tenant data in order to run the service, diagnose faults and answer support requests — our
Privacy Policy says exactly what that covers and what it does not.
Is my cloud provider credentials stored securely?+
Credentials required for platform connections (Azure service principal secrets, AWS access keys, GCP service-account keys, MuleSoft client secrets, Kubernetes tokens, Docker registry tokens, Microsoft 365 app credentials, etc.) are encrypted at rest using AES-256-GCM before storage, never written to logs, and excluded from exports and the API. Where we offer a way to connect without a shared secret, such as the secretless federated Azure connection (Beta), we recommend that path. You can revoke our access from your own cloud provider at any time.
Does CloudMonitor have write access to any of my infrastructure?+
No. All platform connections use read-only access. CloudMonitor cannot create, modify, or delete any resources in Azure, Microsoft 365, AWS, GCP, Kubernetes, Docker, MuleSoft, or any other connected platform.
Is two-factor authentication supported?+
Sign-in is handled by our central identity service at auth.peripatetical.com using a one-time link sent to your email address — there is no CloudMonitor password to steal or reuse. Each person can also turn on two-step sign-in in their account at auth.peripatetical.com: after the email link, sign-in then also asks for a passkey or a code from an authenticator app, so someone who gets into your email still cannot get into CloudMonitor. It is optional for each person; an account admin cannot yet require it for the whole team. On every band, an account admin can set up single sign-on with your own identity provider over OpenID Connect or SAML 2.0 (Google Workspace, Microsoft Entra ID, Okta and others that support them): your people then sign in through it, under its multi-factor rules, and once a test sign-in works you can make it the only way in for your domains. SAML is tested against Keycloak;
tell us if you run something else and we will test it with you. There is no SCIM user provisioning.
Is CloudMonitor SOC 2 compliant?+
No. We have assessed ourselves against SOC 2 and ISO 27001, but no auditor has examined us and we hold no report or certificate. Every change goes through a documented PR flow with audit logging, encrypted-at-rest customer data, one central sign-in with per-tenant data isolation, and an internal living checklist of every AICPA Trust Services Criterion. We have not yet pursued the formal audit because we are pre-revenue; we will start a SOC 2 Type 1 the moment a customer contract makes it economic. If SOC 2 is a procurement blocker,
contact us — we will be straight about timing and can sign contingent on pursuing certification.
⚡ Billing & plans
Pricing, trials, upgrades, and refunds.
What plans are available?+
One product, sold in four sizes. You are billed on
units: a core resource (a VM, database, Kubernetes cluster, app service, load balancer) is 1 unit and a supporting one (a secret, certificate, endpoint, bucket, disk, DNS zone, security group, public IP) is 0.2, so five supporting resources make one unit; the total is rounded up. The band your units fall in is a flat monthly price:
Starter $29 up to 25 units,
Team $99 up to 100,
Business $299 up to 500,
Scale $749 up to 2,000. Example: 20 VMs plus 100 supporting resources is 40 units — Team, $99/mo. Every scanner and alert channel is included at every size and users are unlimited; the REST API is included in every band and AI Insights from Team up. Three optional add-ons (an MSP bundle, a compliance evidence pack, and AI Insights on Starter) are priced separately. Above 2,000 units we quote you directly.
See the bands and the unit calculator →
Is there a free trial?+
Yes — a new account starts with a 14-day free trial. No credit card is required to start, and the trial does not turn into a paid subscription by itself.
What happens when my trial ends?+
Nothing is charged. The trial never converts into a paid subscription on its own — you only pay if you choose to subscribe. To continue past your trial, choose a plan on the Billing page of your dashboard.
Can I change my plan?+
There is nothing to change by hand: the bands are sizes of one product, not separate plans, and every band has every feature. Your band follows your unit count — if it has grown past the top of your band when the billing period closes, that invoice is at the next band's price; if it has shrunk back, it is at the smaller one. No upgrade or downgrade step.
Do you offer annual billing?+
Yes. Paying yearly costs ten times the monthly price, so two months are free. The yearly option sits beside the monthly one on your billing page. The
Terms cover what happens if you cancel or outgrow your band during the year.
What is the refund policy?+
You can cancel at any time; cancellation takes effect at the end of the paid period and there is no termination fee. If the service materially fails to do what we said it would within the first 30 days of a subscription, open a support ticket and we will arrange a refund of the relevant fee. Beyond that, fees already paid are non-refundable except where the law or our payment provider's policies say otherwise. The
Terms are the authoritative version.
Are there per-resource fees or seat charges?+
Nothing is charged per resource, and there are no seat charges. Each scanned item counts toward your
unit total — 1 unit for a core resource such as a VM, database or cluster, 0.2 for a supporting one such as a secret, endpoint or bucket — and the band that total falls in is one flat monthly price; the
pricing page lists every kind we count and how it is weighted. Users are unlimited, and there are no per-platform or per-connection surcharges. The count is measured by the service, and each month's band is set by the most recent unit count in that billing period, so removing resources before the period ends can lower the bill.
Can I get a custom package built for my needs?+
Yes. Submit a
✨ Custom Package request from
Help → Support (
/index.php?view=support) and describe what you need: which platforms, volume and alert channels. It arrives as a support ticket; we reply with a scope and a price benchmarked against our published bands (see
/pricing), and nothing is built until you have agreed both in writing. Same gate applies as for connector requests: we won't take on custom work supporting illegal activity or services with a clear malicious-use risk.
▲ Technical questions
Scan intervals, APIs, integrations, and other details.
Is there an API?+
Yes, in every plan. An account administrator can create keys so your own tools can pull findings, configuration and alerts. A key reads only, unless the administrator allows it to acknowledge, snooze and dismiss findings when creating it. Keys always expire and are rate-limited. The API documentation is in the dashboard once you are signed in.
What counts as an "endpoint"?+
Any HTTP or HTTPS URL you want CloudMonitor to monitor for uptime, response time, and SSL certificate expiry — internal services, customer-facing websites, APIs, health-check URLs. Each unique URL is one supporting resource — 0.2 of a unit, so five endpoints make one unit.
Can CloudMonitor monitor internal endpoints not accessible from the internet?+
Currently, endpoint monitoring requires the URL to be reachable from CloudMonitor's scanner infrastructure. For internal-only services, you can expose a health-check endpoint via a private tunnel or firewall rule. Native private endpoint support is on the roadmap.
How often does CloudMonitor scan my infrastructure?+
Endpoint uptime and certificate checks run every 5 minutes on a paid plan (every 15 minutes during a trial), as a scheduled job that does not depend on anyone being signed in. Each endpoint is probed from every resolver you have configured, and each vantage point alerts and clears independently, so a backend that is down behind one IP and healthy behind the others is reported as exactly that. Any failed check is repeated about 20 seconds later and only a second failure is recorded as down, so a single blip does not raise an alert; the alert for a confirmed failure is dispatched in the same execution as the check that found it. Cloud scans are split by how fast the underlying thing can change. Findings the provider has already raised — threat findings from Defender, GuardDuty, Security Hub and Security Command Center, high-risk Azure Activity Log changes, outages from Azure Service Health, AWS Health and Google Cloud Service Health, and connected Statuspage incidents — are read every 5 minutes on a paid plan (every 15 minutes during a trial). Exposure and state — public network access on storage, SQL and Cosmos DB, security-group and NSG rules open to the internet, unauthenticated container ingress, registry admin users, cluster and node state — is re-checked every hour on Azure, AWS, Google Cloud and Kubernetes. Both dispatch their own alerts in the same run rather than waiting for a separate delivery pass. Posture and expiry — Key Vault and ACM certificates, secrets and access keys, domain expiry, MuleSoft runtime versions, end-of-life dates — runs once a night: its warning ladder steps in days (90/60/30/14/7/1), so a finer cadence would change nothing you receive. Apart from the 5-minute checks, which come with a paid plan, these cadences apply on every CloudMonitor plan; some sources also need a service or support plan from your cloud provider (see “Do any checks need a paid plan from my cloud provider?”).
Can I use CloudMonitor alongside my existing monitoring tools?+
Yes. CloudMonitor is designed to complement tools like Azure Monitor, Amazon CloudWatch, Datadog, Grafana, and PagerDuty — not replace them. It fills the gaps those tools miss: expiring secrets, orphaned storage, and idle VM cost. Alerts can be sent to a Slack incoming webhook or a webhook URL of your own, and the optional API lets your tools pull findings.
Does CloudMonitor support Docker?+
Yes, on every plan. Connect a container registry — Docker Hub, GitHub Container Registry, or any self-hosted registry that speaks the Docker Registry v2 API (registry:2, Harbor, Nexus, Artifactory, GitLab, Quay) — and we flag images anyone can pull,
:latest tags older than a year, and access tokens with no expiry. Connect a Docker Engine host over TCP with mutual TLS and we flag containers running as root, privileged or host-network containers, restart loops, sensitive ports published to the network, unbounded logs and images older than 90 days. Exposure checks run hourly and everything runs nightly, read-only. Two limits: hosts must be reachable on a public address (we cannot scan a daemon that is only on a private network or a local Unix socket), and Docker Hub, GHCR and plain Registry v2 publish no vulnerability results, so we say so rather than call an image clean. Docker Hub and GHCR connections are in Beta: not yet validated on a live customer account.
Setup steps.
What about AWS and GCP support?+
GCP scanning is live. AWS scanning is in Beta (verified against emulated AWS APIs; not yet validated on a live customer account). Both are included on every plan. AWS coverage spans 27 service groups, including EC2, RDS, S3, IAM, Lambda, ACM, Secrets Manager, GuardDuty and Security Hub, load balancers, Route 53, CloudTrail, backup and cost. GCP coverage spans 25, including Compute Engine, Cloud SQL, Cloud Storage, BigQuery, IAM, Secret Manager, Cloud Run, Cloud DNS and Service Health. Each scanned resource adds to your unit count exactly as an Azure resource does — 1 unit for a core resource, 0.2 for a supporting one.
Can I export findings data?+
Yes, in two ways. An account admin can download everything in the account as one ZIP file from Settings › Your data: a JSON file for each data set plus a manifest that lists them and says what was left out. And with the REST API, included in every plan, your own software can pull findings, configuration and alerts. Stored credentials are never included in either.
Do any checks need a paid plan from my cloud provider?+
A few do. CloudMonitor reads what your providers already produce; it never turns a service on or buys anything on your behalf, and none of these costs are part of your CloudMonitor bill. AWS Health (outages in your regions, scheduled changes and retirements) works only on a Business- or Enterprise-level AWS Support plan: Business Support+, Enterprise Support or Unified Operations, or the older Business, Enterprise On-Ramp and Enterprise plans. GuardDuty and Security Hub findings appear only where you have turned those AWS services on, and AWS bills for them. Google Cloud Security Command Center findings are opt-in and need the paid Premium or Enterprise tier. Google Cloud Service Health needs the Service Health API enabled and the Service Health Viewer role; the connect page shows the commands. Azure Service Health and Defender for Cloud’s Secure Score need no extra plan. When AWS Health or Google Cloud Service Health cannot be read, its findings page shows one notice saying why, instead of an empty list.
Is there an iPhone or Android app?+
Not yet — CloudMonitor runs in any mobile browser, and alerts reach you by email, Slack or webhook. A native iPhone and Android app can be developed on request with about 3 months of lead time. Ask via
Help → Support (
/index.php?view=support), or the public
Contact page if you don't have an account.