Every way your service can quietly fail, checked every 30 seconds.
Nine kinds of checks from several regions at once, confirmed before they alert. When something does fail, the incident already knows which deploy came first.
Nine checks, one place to see them.
Websites and APIs, the ports behind them, the records and certificates in front of them, and the jobs that run when nobody’s looking.
-
HTTP & API
Any method, headers and body. Accept a status range like
200-299, follow redirects or not, and mark responses slower than your threshold as degraded. -
Keyword
An HTTP check that also reads the body: down when a keyword is missing, or when an error string shows up. Case-sensitive if you need it.
-
TCP
Connects to any host and port, with or without TLS. Databases, mail servers, brokers: anything that listens.
-
Ping
Round-trip time to a host. ICMP isn’t available from our edge network, so ping measures a TCP connect to the port you choose (443 by default).
-
DNS
Resolves A, AAAA, CNAME, MX, TXT or NS records over DNS-over-HTTPS and checks the answer matches what you expect.
-
SSL expiry
Validates the certificate and its expiry, degraded days before it runs out (14 by default) and down on an invalid certificate.
-
Domain expiry
Reads your registration’s expiry over RDAP and warns 30 days ahead, before a lapsed card takes the domain with it.
-
Heartbeat
Your job pings a URL. If the ping doesn’t arrive within its period plus a grace window, the monitor goes down. A
/failping says so right away. -
Cron
Like a heartbeat, but the expected runs follow a cron expression in your time zone, so a 02:00 job is late at 02:05, not tomorrow.
One flaky hop never pages your team at 3am.
Each monitor checks from the regions you choose: US East, EU West, Asia Pacific by default, 7 in all. A monitor goes down only when enough regions fail on their latest check and the failure repeats on an immediate re-check. Recovery needs a majority of regions back up.
- Pick up to 7 regions per monitor: US East, US West, EU West, EU East, Asia Pacific, Oceania, South America
- Set how many regions must agree before it counts (
confirmations, 2 by default) - Every result records the data center it actually ran from
- US East us-east Failed: timeout after 10s
- EU West eu-west Failed: 503 Service Unavailable
- Asia Pacific ap-south Passed: 200 OK · 231ms
- 2 of 3 regions failed: re-check scheduled immediately
- Re-check fails in the same 2 regions
- Down Incident opened, on-call paged
See the slow slide before the outage.
Not every problem is a 500. Pulseback has a degraded state for the things that are about to hurt, so you can fix them in daylight.
- Responses slower than your threshold mark the monitor degraded
- Certificates go degraded days before expiry, down at expiry
- Domains warn a month before the registration lapses
- Keyword checks catch a 200 OK that’s really an error page
The evidence is already there when you need it.
Uptime is derived from confirmed outages, not from single failed checks, so the numbers you show customers are the numbers that happened.
-
90-day uptime bars
Daily uptime from confirmed outages, the same bars your status page shows. Kept for 13 months on every plan.
-
p50 and p95 by region
Hourly response-time percentiles per region for 90 days, so “slow in Europe” is a chart, not a hunch.
-
Every raw result
Status, timing and the region each check really ran from, kept for your plan’s retention window.
-
Maintenance windows
Schedule planned work and alerts are suppressed for the monitors involved. Subscribers hear about it in advance.
-
Test now
Run a monitor from every region on demand and see each result before you save it.
-
Deploys on the timeline
Connect GitHub or a deploy webhook and the deploy that probably caused a failure is attached to the incident.
Fastest check interval
Compare plans- Free
- 5 minutes
- Starter
- 60 seconds
- Team
- 30 seconds
- Business
- 30 seconds
- Enterprise
- 30 seconds
Raw check results kept
Compare plans- Free
- 7 days
- Starter
- 30 days
- Team
- 90 days
- Business
- 13 months
- Enterprise
- 13 months
Daily uptime is kept for 13 months on every plan.
Monitoring as code, from day one.
Everything you can click is in the REST API, with workspace API keys. Tag monitors by environment, and review changes like any other pull request with the Terraform provider.
Terraform provider docs# monitors.tf resource "pulseback_monitor" "checkout" { url = "https://checkout.acme.dev/health" interval = 30 regions = ["us-east", "eu-west", "ap-south"] confirm = 2 escalation_policy = pulseback_policy.platform.id }
Bring your monitors over in one step.
Import with a StatusCake API key or a CSV export. Pulseback creates the monitors, tells you which ones it skipped and why, and you can run both side by side until you trust it.
Moving from StatusCakeGet paged when Pulseback launches.
Pulseback is almost ready. Join the waitlist and we’ll email you the day it opens, Free plan and all.
Join the waitlist