Review build Not the live site — unlaunched products, unfinished copy, no prices. comsky.ai →
Comsky Monitor
Join the waitlist
Coming soon

Know it broke before your customer tells you.

Website and API uptime, server and VM agents, SNMP network devices and log alerts on one console — with escalation to the person actually on call, over WhatsApp, SMS or voice.

In build through 2026 · Checks from Indian locations first · Alerts over WhatsApp, SMS and voice · One account with Comsky Cloud and Backup

STATUS BOARD02:14 IST
comsky.in up · 214 ms
api.comsky.in/health up · 96 ms
ssl · comsky.in 11 days left
vm-prod-02 · disk 94 % used
core-sw · gi1/0/2 down 4 min
nightly-backup no check-in · 26 h
WHATSAPP SENT TO OPS · VOICE CALL IN 10 MIN
The part everyone else treats as a footnote

At 2am, an email is not an alert.

Every monitoring tool can tell you a site is down. The question is whether that fact reaches a person who is asleep. Comsky Monitor will send the first alert over WhatsApp, because in India that is the message that gets read — then SMS, then a voice call that keeps ringing.

ESCALATION, ONE INCIDENTNOBODY HAS ACKNOWLEDGED
0 min WhatsApp to whoever is on call Acknowledge and Resolve on the message itself
5 min SMS on a DLT-registered template Only if nobody has acknowledged
10 min Voice call through Comsky Telephony A phone that rings until it is answered
30 min The team lead, then the account owner The chain keeps going until someone takes it
ACKNOWLEDGE ANYWHERE IN THE CHAIN AND THE REST OF IT STOPS

WhatsApp, with buttons that work

The alert arrives as a WhatsApp message with Acknowledge and Resolve on it. Tapping Acknowledge stops the escalation chain and writes who took the incident, without anyone opening a console.

SMS on DLT-registered templates

Transactional SMS in India needs templates registered under TRAI’s DLT regime. Comsky will handle that registration, so alert SMS is not the thing that quietly stops delivering six weeks after you set it up.

A voice call that rings

Escalation to a phone call placed through Comsky Telephony, on Comsky’s own carrier legs rather than a reseller’s. A ringing phone is the only channel that reliably wakes someone.

On-call rosters, not a mailing list

Shift-based rosters decide who is paged, with per-channel rules by time of day — email during working hours, WhatsApp and voice between 6pm and 8am.

Maintenance windows

Schedule a window and checks keep running while alerts stay quiet. The downtime inside the window is excluded from the SLA figures, so a planned migration does not damage the monthly report.

Confirmed before it pages

A failing check is re-tested from a second location before anyone is woken. A flaky route from one city should not cost your engineer a night.

Slack, Microsoft Teams, Telegram, email and plain webhooks will be there too. They are just not the channel we are designing around.

Coverage

One console for the website, the server, the switch and the backup.

The reason most businesses run no monitoring at all is that doing it properly means three tools and three bills. This is the set an Indian SMB or MSP actually needs, in one place.

MonitorWhat it looks atIt alerts when
WebsiteStatus code, response time, and a keyword that must appear on the pageThe code is wrong, the page is slow, or the keyword is gone
REST APIA request with your headers and auth token, and an assertion on the bodyThe assertion fails or the endpoint times out
Port and pingA TCP port and ICMP reachabilityThe port refuses or the host stops answering
DNSThat a record still resolves to the value you expectThe record changes or the resolver fails
SSL certificateChain validity and days remainingThe chain breaks, or expiry is closer than your notice period
DomainRegistration expiryThe domain is close to lapsing
HeartbeatAn inbound URL your cron job or backup has to pingThe ping does not arrive inside its window
Server agentCPU, memory, disk, load, processes and services on a machineA threshold is crossed or a service stops
SNMP deviceRouters, switches, firewalls, access points and UPS unitsAn interface drops or a device stops responding
Synthetic journeyA scripted browser run of login, add to cart and checkoutAny step in the journey fails
Log ruleA pattern in your logs that should never appearThe pattern appears, or stops appearing

Checked from here first

MumbaiDelhi NCRBengaluruChennaiHyderabadKolkataPuneSingaporeFrankfurtLondonVirginia

The count of global locations is the wrong number to compete on. What matters to a business in Noida is whether the check ran from Mumbai on an Indian route, and whether a second Indian location agreed before anyone was woken. The list above is the planned first set and is being built out.

Why it comes from Comsky

It will already know what you are running.

A standalone monitoring tool starts from nothing and asks you to describe your estate. Comsky Monitor will start from an account that already contains your machines and your backup jobs.

01

A Comsky Cloud VM enrols itself

Create an instance in Comsky Cloud and the agent is already on it, with a default monitor set applied. There is no install step and no copied API key, because both sides are the same account.

  • CPU, memory, disk and load from first boot
  • Instance destroyed → monitor closed, not left alerting
02

Backup jobs check in like any other monitor

A Comsky Backup job that finishes reports in. One that does not is an incident. “No check-in for 26 hours” pages exactly like a down website, which is the alert almost nobody has today.

  • Missed schedule, failed job and stalled agent are separate alerts
  • The failure is found on a Tuesday, not at the restore
03

Everything else you add by hand

Machines that are not on Comsky Cloud take the same agent. Network devices are discovered by scanning an IP range. Websites and APIs need a URL. Nothing here is locked to our infrastructure.

  • Windows, Ubuntu, Debian, RHEL, Rocky and AlmaLinux
  • SNMP discovery across a CIDR block
04

One pipeline, one bill

A dead VM, a saturated branch link, a missed backup and a down checkout page all arrive through the same escalation policy, on the same invoice, from a support desk in the same time zone as you.

  • One on-call roster covers the whole estate
  • GST-inclusive rupee invoicing alongside the rest of Comsky
Status pages and SLA

Proof you were up, on a page your customers can read.

An uptime number is worth very little when you are the only one who can see it. A public status page answers the support queue before it forms, and the SLA report is the artefact an MSP attaches to a monthly invoice.

ALL SYSTEMS OPERATIONALSTATUS.YOURCOMPANY.IN
Website99.9871 %
Customer API99.9412 %
Payments100.0000 %
90 DAYS AGOTODAY

Your domain, your logo

A public or password-protected page on status.yourcompany.in with a certificate issued automatically. Components grouped the way your customers think about your service, with per-component email and RSS subscriptions.

SLA reports to four decimals

Define the target per service and the report computes against it — 99.9871 %, not “about three nines”. Scheduled maintenance is excluded from the maths, and the incident timeline sits under the number.

An MSP console, not a checkbox

One login across every client you manage, with hard isolation between tenants, your name and colours on the portal and the reports, and engineer roles that limit who can see whose estate. Priced per managed customer rather than per host.

A poller inside the client LAN

Most Indian clients will not open a firewall to a SaaS. A small on-premise poller watches internal apps, printers, NAS boxes and LAN devices and pushes results out, so nothing internal is ever exposed to the internet.

Everything else

The rest of it, without the paragraphs.

Planned for the first release unless the line says otherwise.

Uptime checks

  • HTTP and HTTPS status and keyword
  • REST API with headers, auth and body assertions
  • TCP port and ICMP ping
  • DNS record verification
  • SSL chain and expiry
  • Domain registration expiry
  • Inbound heartbeat URLs for cron and backups

Servers and VMs

  • Agent for Windows and Ubuntu, Debian, RHEL, Rocky, AlmaLinux
  • CPU, memory, load and swap
  • Disk space, IOPS and latency
  • Network throughput and interface errors
  • Process and service checks with a top-process view
  • Custom plugins in Python, Shell or PowerShell
  • Docker host basics

Network

  • SNMP polling for routers, switches, firewalls and UPS
  • Discovery by IP range or CIDR block
  • Per-interface bandwidth, errors and latency
  • SNMP trap receiving
  • Device templates by vendor and model
  • On-premise poller behind the firewall

Logs and experience

  • Agent log tailing and syslog receive
  • Automatic grouping of similar messages
  • Alert rules on a log pattern
  • 7, 15 and 30-day retention tiers
  • Scripted browser journeys on a schedule
  • Core Web Vitals and JS errors by city and ISP (later release)

Alerting and on-call

  • WhatsApp with Acknowledge and Resolve
  • SMS on DLT-registered templates
  • Voice call through Comsky Telephony
  • Slack, Teams, Telegram, email and webhooks
  • Time-ordered escalation chains
  • Shift rosters with per-channel time rules
  • Maintenance windows that suppress and exclude

Reporting and access

  • Public and private status pages on your own domain
  • SLA reports to four decimal places
  • Per-component subscriber notifications
  • Dashboards you arrange yourself
  • Role-based operator access
  • Per-tenant isolation for MSP accounts
  • Data held in the Indian zone
Why Comsky Monitor

Against the incumbent, and against doing nothing.

Site24x7 is the closest analogue and it is a good product. The more common competitor is no monitoring at all — finding out when a customer rings.

Site24x7No monitoringComsky Monitor
How you find outEmail, SMS, voice, mobile push and 30+ integrationsA phone call from an annoyed customerWhatsApp first, then SMS, then a voice call — with Acknowledge on the message itself.
WhatsApp as a channelNot a native channelThe customer’s channel, not yoursNative, and the default first hop
Who gets wokenOn-call rosters and escalation profilesWhoever happens to pick upRosters and time-ordered escalation, included rather than reserved for a higher tier.
SMS and voice legsMetered alert credits through their own providersComsky Telephony will carry them, with DLT template registration handled for you.
Your Comsky Cloud VMsAnother host to install an agent onEnrolled with the agent at the moment the instance is created.
A missed backupOnly if you wire up a heartbeat yourselfYou find out at the restoreA first-class monitor. Comsky Backup jobs page like a down website.
Where the data sitsIndia is one of several regions you chooseNowhere — there is no history to look atThe Indian zone, the same one Comsky Cloud runs in.
What you are billed inUSD by default, INR behind a currency selectorNothing, until the outage costs you somethingRupees, GST inclusive, on the same invoice as the rest of Comsky.
Surface areaAPM, tracing, RUM, session replay, Kubernetes, cloud costUptime, servers, network, logs and alerting. Narrower on purpose.
Available todayYes, for yearsAlwaysNot yet. In build, with first accounts through 2026.

If you already run Site24x7 or UptimeRobot and the alerts reach someone, there is no reason to move yet. The case is stronger when your machines and your backups are already on Comsky, or when the honest answer to “who gets the alert at 2am” is nobody.

Pricing

Three editions. The numbers are not set yet.

The shape is decided: a free tier that is genuinely useful for uptime, one paid tier that covers servers and networks, and a separate MSP edition priced per managed customer rather than per host.

Uptime

On request

Websites, APIs, certificates and heartbeats. Enough for a small team to run on its own.

  • Website, API, port, ping and DNS checks
  • SSL and domain expiry warnings
  • Inbound heartbeat URLs
  • One-minute intervals
  • Second-location confirmation before alerting
  • WhatsApp and email alerts
  • One public status page
Get early access

Infrastructure

On request

Servers, VMs and network devices beside the websites. The edition most businesses will buy.

  • Everything in Uptime
  • Server and VM agent on Windows and Linux
  • Process, service and custom plugin checks
  • SNMP devices with IP-range discovery
  • Log collection with alert rules
  • Escalation chains and on-call rosters
  • Voice call escalation
  • Comsky Cloud auto-enrolment and Comsky Backup job monitors
Get early access

MSP

On request

One console for every client you look after, with your name on the portal and the reports.

  • Everything in Infrastructure, per tenant
  • Multi-tenant console with hard isolation
  • White-label portal and status pages
  • On-premise poller inside the client LAN
  • Per-client SLA reports to four decimals
  • Granular engineer roles
  • One licence pool across tenants
Talk to us

Comsky Monitor is still being built, so no price is published. It is not decided yet, rather than decided and withheld. A free tier for uptime monitoring is planned, and alert credits for WhatsApp, SMS and voice are intended to be bundled generously rather than metered — Comsky owns those pipes. Tell us what you need to watch and we will price it against your actual estate when the numbers are set.

Questions

Before you put your name on the waitlist.

Can I use Comsky Monitor today?

No. It is in build, and first accounts are going out through 2026. Everything on this page is what the product is being built to do, not what it does this morning. If a date matters to you, ask us and we will tell you where the piece you need actually is.

Will alerts really come over WhatsApp?

That is the part we are designing around. Alerts go out through the WhatsApp Business API that Comsky Chat already uses, with Acknowledge and Resolve buttons on the message. SMS runs on templates registered under TRAI’s DLT regime, and voice calls are placed through Comsky Telephony on our own carrier legs — so the SMS and voice margin a monitoring vendor usually resells sits with us instead.

Do I have to open my firewall to you?

No. For anything internal, a small poller runs inside your network, performs the checks locally and pushes results out over an outbound connection. Nothing internal is exposed to the internet and no inbound port is opened. This is also what makes the MSP edition sellable to clients who will not permit inbound access.

Does it only work with Comsky Cloud?

No. The agent installs on any Windows or Linux machine, wherever it runs — your office, a rack you own, or another provider. The Comsky Cloud tie is a convenience: instances created there arrive already enrolled, with a default monitor set. Everything else you add the normal way.

What does Comsky Backup integration actually mean?

A backup job becomes a monitored object. A job that finishes reports in; a job that does not raises an incident and runs the same escalation chain as a down website. The alert reads “no check-in for 26 hours”, which is a thing most businesses discover only when they try to restore.

Will it do APM, tracing, session replay or Kubernetes?

Not at launch, and we would rather say so than imply otherwise. Those are where Datadog and Site24x7 spend their engineering, and they are not what an Indian SMB or MSP is missing. OpenTelemetry ingest is on the list as the way in for teams that already instrument their applications, but it is not built yet.

Where will the monitoring data be held?

In the Indian zone, the same one Comsky Cloud runs in. Check results, agent metrics, logs and incident history are created there and stay there. Checks themselves run from monitoring locations in several countries, because that is the only way to tell a real outage from a bad route.

What happens during planned maintenance?

You schedule a window. Checks keep running so you can still see what is happening, alerts are suppressed, and the downtime inside the window is excluded from SLA calculations and from the public status page history. A migration on a Sunday night should not cost you a nine.

Start with Comsky Monitor.
Grow into the ecosystem.

One account opens every Comsky product — and one invoice covers them, however many you run.