SLO tracking with error budgets
Set a target per service, such as 99.9%, and see how many minutes of downtime you can still afford this month, not just an uptime percentage.
The uptime monitor that tells you why.
Uptime monitoring for developers and small teams running websites, APIs and cron jobs. Every alert says what failed, where, and whether it looks like your fault, a vendor’s or the network’s.
50 monitors free forever, no card.
Illustration, not live data: monitoring locations check one website. One region sees a timeout and retries once, other regions see the same failure on their own checks, the failure is confirmed, the on-call team is alerted, and the site recovers.
Alerts go where your team already is
- Slack
- Discord
- Telegram
- Microsoft Teams
- Google Chat
- Mattermost
- Push notifications (browser)
- SMS
- Voice calls
- Webhooks
- PagerDuty
- Opsgenie
An alert that explains itself
“Site down” is where the work starts. SutramX attaches the reasoning to every incident and puts a one-line verdict in the alert itself, so the first question is already answered.
Down from every region: timeout (all regions agree), likely not your fault: Stripe API degraded for several SutramX customers.
- Regions
- Frankfurt: timeout · Arizona: timeout (2 of 2 agree)
- Failure
- Timeout calling api.stripe.com, all regions
- Alert decision
- Alert sent: quorum met, no maintenance window
- Official status
- Not on Stripe’s status page yet
- Why this alertEach incident shows every region’s result, the quorum rule applied, the failure class (DNS, TLS, timeout, 5xx…) and why an alert was or was not sent. Plain rules, no AI, on every plan.
- Vendor outage detectionWhen Stripe, OpenAI or an AWS region fails for several SutramX customers at once, your incident says “likely external”, from our own checks rather than the vendor’s status page.
- Deploy correlationSend deploy events from GitHub, GitLab or any CI, and an incident that starts soon after a release names the deploy that likely caused it. On Growth and Pro.
- Flakiness scoreA 0–100 score per monitor over 7 and 30 days, with the top reasons, so you know which alerts to trust.
What happens when something breaks
Follow one outage through 5 steps, from the first failed check to the report you read the next morning.
A check fails
Each monitor runs on its own schedule, as often as every 15 seconds. A timeout, a wrong status code, a missing keyword or a failed API assertion all count as a failure.
The failure is confirmed
Before anyone is paged, the failed check is retried once from the same region and has to keep failing for the monitor's failure threshold. Paid plans check from more than one region, and enough of those regions must also see it fail. A problem on one network path stays a note in the log instead of a 3 a.m. alert.
The right people hear about it
One incident sends one alert, to the channels your team already watches. Escalation policies and on-call schedules decide who is woken up, and when to go further if nobody acknowledges within, say, 5 minutes.
Customers see what is happening
Post updates to a public status page, on your own domain from the Growth plan. Deployment correlation shows whether the last release lines up with the first failure, and monitor dependencies keep one database outage from firing 50 alerts.
You see what it did to your SLO
SLO tracking shows how much of this month’s error budget the incident used and how much is left, so the next reliability decision starts from a number. For a 99.9% monthly target that budget is about 43 minutes.
9 kinds of monitor. 14 alert channels. Checks as often as every 15 seconds.
Confirmed from more than one network before anyone is paged
Checks from three locations reach your endpoints over different networks. When one of them sees a failure, the others are asked, and on paid plans at least 2 must agree, before anyone is woken up.
- Frankfurt
- Arizona
- Mumbai
Nine kinds of monitor, two expiry reminders
From a single landing page to the nightly backup job nobody remembers until it stops. All 9 are on every plan, including Free.
- HTTP and HTTPSStatus codes, response times and redirects for any URL.
- API with assertionsCheck JSON fields, headers and status codes in every response.
- KeywordFail when a page loses the text it should show, or shows text it should not.
- PingReachability and round-trip time for hosts and network gear.
- TCP portConfirm databases, mail servers and other services accept connections.
- UDP portDNS servers, game servers, VoIP and anything else that answers over UDP.
- Cron and heartbeatYour job checks in on schedule. Silence opens an incident.
- DNS recordsA, AAAA, CNAME, MX, TXT and NS records: alert when they change or stop matching what you expect.
- MCP serversHandshake, protocol version and tool list of remote MCP servers, with an alert when the tools change.
Plus two expiry reminders
- SSL expiryReminders on HTTP and API monitors, well before a certificate runs out.
- Domain expiryReminders before the domain behind a monitor lapses.
Built for the whole incident, not just the ping
Knowing a site is down is the start. These are the parts that shorten the outage and explain it afterwards, with 90 days of raw check history on every plan.
Deployment correlation
See at a glance whether an incident started right after a release, and roll back sooner. Included from Growth ($29/month). Included from Growth (₹1,499/month).
Escalation policies
If nobody acknowledges, the alert moves up the chain on the schedule you set, for example after 5 and then 15 minutes.
Monitor dependencies
When the database fails, you get 1 alert about the database, not 50 about everything built on it.
Status pages on your domain
90 days of uptime history and incident updates at your own address, with email updates for subscribers.
On-call schedules
Rotations decide who is paged tonight, so alerts reach whoever is actually on duty. Included from Pro ($59/month). Included from Pro (₹2,999/month).
Priced in rupees
Teams in India pay in INR: Starter is ₹599/month, Growth ₹1,499/month and Pro ₹2,999/month, all incl. GST.
Two months free on annual
Pay monthly, or pay annually and get 12 months for the price of 10. Your price stays locked while you stay subscribed.
Unlimited seats
Invite the whole team on every plan, Free included, at $0₹0 per seat. Nobody has to share a login to see an incident.
Every region on one screen
Response times, daily uptime and the results from up to 3 regions behind each check, side by side. Try a real check of your own site from every region below.
Multi-region check
Free, no signupSends one real request from each monitoring location and shows the status code and response time each one got.
Frankfurt
Arizona
Mumbai
Go deeper
Guides, free tools and setup pages for the questions that come after “is it up?”.
- Uptime monitoring guidesPractical answers, no signup needed.24 guides
- Free reliability toolsCheck a site or do the maths, no account needed.16 tools
- Alert integrationsStep-by-step setup for the tools your team already uses.11 setup guides
- Compare SutramXFair, dated comparisons, including where others are stronger.5 comparisons
- Monitoring use casesWhat to watch, and how often.3 use cases
A single failing check is a hypothesis. Agreement across regions is an incident.
Start free. Pay when it matters.
Four plans, from $12/month after Free,, from ₹599/month after Free, and one conversation for larger teams. Compare every plan
Recommended
Growth
$29/month incl. taxes₹1,499/month incl. GST
200 monitors, checked every 30 seconds.
Get GrowthStart free — Free plan forever, no card required. Upgrade any time. Every plan includes unlimited team seats. All prices include taxes — nothing extra at checkout.
Know it’s down before your customers do.
Start free — Free plan forever with 50 monitors, no card required. Upgrade any time.