HTTP(S)
Checks the expected status (200, 301, etc.) within the timeout. On HTTPS targets, watches the SSL certificate and warns 30 days before expiry.
Homepage, REST API, admin panel, payment webhook.
Sentinela Uptime
Know it went down before your customer does.
Checks from our infrastructure in Brazil, with a configurable interval from 1 minute. Seven observation modes covering HTTP, keyword, ping, port, push, DNS and Vercel error logs. Alerts in seconds via the channels you choose.
No card to start · 5 monitors free forever · 2-minute setup
Check types
Seven modes, one for each case.
Six in pull mode (Sentinela fetches the target) and one in push mode (the target pings us — for internal jobs with no exposed HTTP).
Keyword
Extends HTTP by matching (or negating) a keyword in the response body.
Detect an error page that returns 200, an unscheduled maintenance banner, defacement.
Ping (ICMP)
Sends an echo packet to a host and measures response time.
Server with no exposed HTTP, gateway, network device.
Port (TCP)
Opens a socket on a specific port. For services that don’t speak HTTP.
SSH (22), SMTP (25), MySQL (3306), Redis (6379).
Heartbeat (push)
Inverts the flow: your cron/worker pings a unique URL. No ping within the interval (with 2x grace) — goes DOWN.
Nightly backup, data sync, queue worker, internal job with no exposed HTTP.
DNS (resolve A)
Resolves the host’s A record. UP if ≥1 address comes back. Detects a lost zone and slow propagation without HTTP.
Confirm resolution after an NS change, validate a client’s external zone.
Vercel (error logs)
Queries the official Vercel API on every check. Goes DOWN when the production deploy fails or runtime errors exceed the threshold you set (0 = any new error).
Next.js/serverless app on Vercel: catch a broken deploy or a runtime-exception spike before the customer complains.
Notifications
Seven channels — choose where you get it.
Each monitor can fire to several channels at once. Configure once, reuse everywhere.
Default for everyone. Clean look, direct link to the incident. Multiple recipients per channel.
Webhook
JSON with full payload. Customizable headers. Connect with n8n, Zapier, or your own system.
Telegram
Official bot. Set up in 30s by pasting the chat ID. Great for a phone alert without an extra app.
Slack
Rich message in the team channel, color by severity. Official Slack webhook.
Discord
Colored embed in the server. Ideal for teams already living on Discord.
PagerDuty
Creates a PagerDuty incident when the monitor goes down. Auto-resolves when it comes back. For serious on-call.
Opsgenie
Same model as PagerDuty: creates an alert on downtime, resolves on recovery. For Atlassian teams.
Jira
Opens an issue in your project when the monitor goes down; comments and transitions it to Done when it recovers. Token or OAuth. The same channel turns Security findings into tickets.
The Test channel button fires a real notification, without having to take anything down on purpose. The same channels apply to Security Audit notifications.
Public status page
Transparency without friction.
Every incident tells a story. Your status page tells the whole story — what went down, when it came back, and the uptime over the last 30, 60 or 90 days.
- · Public URL (
status.seudominio.comon higher plans) - · Grouping by service (site, API, panel)
- · Incident history with timeline
- · Uptime% calculated in real time
- · Email subscribers — visitors subscribe and get notified when incidents open/close
- · RSS feed — last 50 events at
/api/status/{slug}/feed.rss - · Optional password protection · toggle to disable subscriptions
- · No login for the visitor
exemplo.com.br
operational
Marketing site
up
Public API
up
Admin panel
up
Payment webhook
up
Sync job
up
Uptime 90 days
99.91% over the last 90 days · 4 incidents
Maintenance windows
A planned deploy doesn’t become an alert.
Going to take the API down at 3 AM Saturday to run a migration? Create a window and:
- · No alerts fired during the window
- · Time doesn’t count against uptime%
- · "Scheduled maintenance" banner on the status page
Auto-pause
Client vanished? Pauses itself.
A client cancelled and nobody told you. The domain has been down for days and your dashboard turns into visual noise. Configure a window in hours — if the incident stays open beyond that, the monitor pauses automatically.
- · Adjustable per monitor (empty = off)
- · Audit trail records the pause
- · Manual reactivation resets the cycle
Latency
p50, p95, p99 — the truth about the slow ones.
Average time lies. When you look at p95, you see how long the tenth-worst user waited. p99 is the worst in 100. Sentinela computes it automatically over all UP checks in the interval and compares week over week with a trend indicator.
- · p50 / p95 / p99 + sample count
- · Weekly trend (↑ worse, ↓ better)
- · CSV export of checks for external analysis
- · Weekly email report + downloadable PDF
api.exemplo.com.br
last 7 days · 10,080 samples
p50
124ms
↓ -8%
p95
487ms
↓ -12%
p99
1.2s
↑ +6%
Samples
10k
Comparison
Sentinela vs alternatives.
| Sentinela | UptimeRobot | StatusCake | Better Stack | |
|---|---|---|---|---|
| Minimum interval | 1 min | 5 min (free) / 1 min | 1 min | 30 s |
| Localized (PT-br native) | ✓ | — | — | — |
| Payment in BRL | ✓ | — | — | — |
| Hosted in Brazil | ✓ | — | — | — |
| Includes security audit | ✓ | — | — | — |
| Public status page | ✓ | ✓ | ✓ | ✓ |
Each product has its strength — pick the one that fits your case. Comparison based on the public plans of the respective platforms.
For technical teams
API, webhook, isolation — as expected.
REST API
Auth via Sanctum. Read monitors, checks, incidents. Business+ plan has write access.
Webhook with retry
Versioned payload. Customizable headers. Exponential retry until confirmation.
Native multi-tenant
Global scope at the database — each user only sees what’s theirs, guaranteed in the query.
2FA TOTP
On every plan. Compatible with Google Authenticator, 1Password, Authy.
Internal audit
Trail of login, lockout, CRUD on monitor/channel — exportable on the Business plan.
Exportable CSV
Check history of each monitor in CSV via the UI or a dedicated endpoint.
FAQ — uptime
Common questions.
How long until a downtime is detected?
Depends on the configured interval. On the Pro plan, a check every 1 min — the alert goes out within 60s after confirmation. We confirm with a retry before opening an incident, to avoid a false alarm from 1 isolated failure.
Where do you check from?
Currently from Brazil (São Paulo). Multi-region is on the roadmap.
Can I monitor a service behind a VPN/firewall?
Not directly — we’re an external observer. But the Heartbeat (push) monitor covers exactly that case: your internal cron pings a unique Sentinela URL, and a missing ping turns DOWN.
Does high-volume webhook cost extra?
No. Webhook has no cost per event. You pay for the plan, not per call.
Can I adjust how many failures open an incident?
Yes. By default we confirm with a retry; on very sensitive monitors you can raise the number of consecutive failures needed to open an incident.
How does the Vercel monitor work?
You connect a read token and the Project ID. On every check we query the official Vercel API: if the production deploy fails or runtime errors exceed the threshold you set (0 = any new error), it opens an incident and fires the same alert channels. The token is encrypted and never displayed after saving.
Also in the same dashboard
Uptime and security in one dashboard — and risk becomes a number.
48 probes across 6 layers (TLS, headers, WAF, malware, WordPress CVE, privacy — LGPD-BR by default and GDPR opt-in, SPF/DMARC, domain lock + optional SAST/secret scanning/Dockerfile/IaC/GitHub Actions/Vercel API), with score 0–100 and grade A–F. <strong>Unique edge:</strong> your real uptime incidents feed the <strong>Financial Cyber Risk Score in money/year</strong> — exposure = observed downtime cost + incident probability × impact. Only possible with uptime and security on the same platform. <a href="/blog/como-calculamos-risco-financeiro-em-reais" class="underline">How we compute the number</a>.
Start with Free
5 monitors, 5-min interval, for free.
No card. Move up to Pro (1-min interval + webhooks) when Free gets tight.