Monitoring features
Check websites, APIs and infrastructure. Configure how failures are confirmed and who receives alerts.
What we can monitor
Checks for websites, APIs, network services, certificates and scheduled jobs.
HTTP/HTTPS
Website and API monitoring with response time tracking, status codes, and content validation
- Response time tracking
- Threshold alerts (Free)
- Status code monitoring
- Content validation
TCP
Port connectivity monitoring for databases, mail servers, and custom TCP services
- Port connectivity
- Service availability
- Network latency
- Connection checks
UDP
Connectionless port monitoring for DNS, VoIP, game servers, and streaming services
- Datagram delivery
- Service availability
- Response validation
- Timeout detection
SSL/TLS
Certificate monitoring with expiration alerts and security analysis
- Certificate expiration
- Chain validation
- Security analysis
- Compliance checks
WebSocket
Real-time connection monitoring for chat apps and live features
- Connection stability
- Message delivery
- Handshake validation
- Standard compliance
gRPC
Microservice monitoring and validation
- Service health
- Method availability
- Response validation
- Performance metrics
MCP
Monitoring for AI Agents and Context Services
- Context availability
- Model connectivity
- Standard compliance
- Response validation
Ping (ICMP)
Network reachability and latency monitoring for hosts and infrastructure
- ICMP echo checks
- Network latency
- Host reachability
- TCP fallback
DNS
Periodic DNS record verification, catch DNS drift, propagation issues, and hijacks early
- A / AAAA / CNAME records
- MX, NS, TXT verification
- Expected-value matching
- Multi-resolver checks
Heartbeat
Inverse monitoring for cron jobs and workers, your job pings us, we alert when it goes silent
- Unique ping URL per job
- Configurable interval + grace
- No auth required on ping
- Alert when a ping misses its grace period
Choose a monitor for your service
Compare plans for supported monitor types and check intervals.
Start monitoringMost alerts are confirmed from at least two regions.
When a check fails, other regions repeat it before we alert you. An alert we cannot confirm is still sent, labelled unverified. Comparing results helps distinguish a service failure from a problem affecting one monitoring location.
- Fixed on Free and Hobby: these plans use the default threshold shown in the table.
- Threshold configurable on Starter+: raise it up to 5 regions on Professional and 6 on Enterprise. Choosing the region a monitor runs from is on Professional (up to 5) and Enterprise (all regions).
- Compare regional results: A failed check from one location alone does not confirm a service outage.
If the second region reaches the site, the alert is still sent, labelled unverified.
The first number is how many regions must see a failure for the alert to say it is confirmed. On Free the second region can confirm an alert but never holds one back. From Hobby up, when most regions reach the site, the alert is held and checked again five minutes later if the first region still sees the failure; any other result is sent, labelled unverified. Free and Hobby use a fixed threshold; Starter, Professional and Enterprise can change it.
Group simultaneous failures into one message.
Related email and Slack alerts are collected into one message with the affected monitors and their shared error. SMS and voice alerts are sent immediately.
- Grouped by what they share: alert type, monitor group and the underlying error. Alerts from different accounts are never grouped together.
- A 60-second grouping window: Alerts collect for 60 seconds so a cascade arrives as one message. The window starts when the first alert lands, a busy group can never delay itself indefinitely.
- Likely-cause hint on Professional and Enterprise: a likely cause and a first step to try, written from the errors your monitors actually returned.
- The alert never waits for the hint: if the hint is slow or unavailable, the grouped alert still goes out, just without it.
Grouping is available on every plan. The likely-cause hint is on Professional and Enterprise.
Alert grouping
+ 37 more
All behind the same upstream, connection refused
Track error budgets and burn rates.
Set an availability target per monitor. UptimeWarden computes an error budget and uses Google SRE's multi-window burn-rate thresholds to determine when to send SLO alerts.
- Budget gauge per monitor: live remaining-budget % and current burn rate, updated every minute by our Go scheduler.
- Two windows must agree. A fast alert fires when the 5-minute burn is 14.4× AND the 1-hour window confirms. A slow alert catches 6-hour / 3-day slow drift.
- Flapping auto-suppressed. If a monitor is bouncing state, burn-rate pages pause until it settles, and the monitor is labelled as suppressed, so a paused alert never looks like a healthy one.
- Same escalation path. Burn-rate events flow through your existing on-call rotations, no new channel to configure.
From failed check to resolved incident
Each failure opens one incident: confirmed from another region, routed to the channels you chose and closed when checks pass again
Detect
Before an alert goes out, the failed check is repeated from another region
Alert
One open incident per monitor and alert type. Alerts go to the channels you picked for each alert type and stay quiet during maintenance windows
Resolve
The incident closes after consecutive successful checks. You can have a postmortem draft written, with a timeline and a likely cause linked to its evidence; every plan includes drafts each month
Detect
Before an alert goes out, the failed check is repeated from another region
Alert
One incident per monitor, sent to the channels you picked, quiet during maintenance
Resolve
Closes after consecutive successful checks. Every plan includes postmortem drafts you can edit
Operational controls, built in
Maintenance windows, on-call, signed webhooks, cascade suppression and postmortems in the same account.
Maintenance windows
Suppress notifications during planned downtime, one-time or recurring weekly. Checks still run, alerts stay quiet.
On-call and escalation
Build weekly rotations, chain escalation steps, and ack with a single click from the email, no login needed.
Signed webhooks (HMAC V2)
Every webhook is signed with HMAC-SHA256 over the timestamp and body. Receivers can reject anything older than five minutes, which blocks replayed requests.
Cascade alert suppression
When an upstream service goes down, downstream incidents are tagged "cascade-suppressed" so each dependent monitor does not send a separate alert.
Postmortem drafts
When an incident closes, a draft postmortem can be written for you: a timeline and a likely cause, each linked to the checks and events it came from. Review and edit it before sharing.
How incidents are opened, alerted and closed
Roles and workspaces for teams
Role-based access control and workspaces, so each team sees the monitors it owns
Role-based access
Who can see and change what
Control who can see, edit, or manage monitors, down to individual resources.
Team workspaces
One per team, project or environment
Create dedicated workspaces for different teams, projects, or environments with isolated monitoring setups.
Team management
Invites by email
Invite team members by email and assign their access levels.
Check regions and your own agents
Checks run from six regions. To check services that are not reachable from the internet, install a lightweight agent inside your own network.
Agents on your servers
Inside your infrastructure
Install lightweight agents on your own servers, containers, or Kubernetes clusters for internal monitoring.
Private networks
Services with no public endpoint
Monitor internal services, databases, and applications that are not exposed to the internet.
Six regions
Checks from several places
Monitor from 6 global regions, so you can tell a local outage from a global one.
What agents and regions are for
Checks from close by
Agents closer to your infrastructure detect issues faster than remote monitoring
Nothing exposed
Monitor private networks without exposing services to the internet
Regional measurements
Compare availability and response times from multiple geographic locations
Runs where you need it
Cloud, on-premise, edge locations or a mix of them
Confirmed incidents
Checks from multiple regions help reduce false alerts caused by temporary network problems.
Run an agent in your network
The agent guide walks through installing it on a server, in Docker or on Kubernetes.
Add your first monitor
Start with the Free plan or compare plans for the features and limits you need.
Talk to us
Questions about plans, integrations or enterprise needs? Send a message, we typically reply within 1 business day.
Company details
Send us a message
Share your questions, partnership ideas, or enterprise needs.