Blind spots
A check from one place sees one route to your site. When a provider in another country cannot reach you, it stays green and you hear about it from a customer.
Product · Uptime monitoring
{ "id": "chk_4b1d9e07a3c2f581", "name": "Cloudflare Edge", "type": "http", "target": "cloudflare.com", "interval_seconds": 60, "state": "up", "uptime_24h": 99.98, "p50_ms": 18, "p95_ms": 42, "locations": ["DE", "US", "SG"], "latest_runs": [ { "country": "DE", "asn": 3320, "status_code": 200, "first_byte_us": 14200 }, { "country": "US", "asn": 7922, "status_code": 200, "first_byte_us": 21800 }, { "country": "SG", "asn": 4755, "status_code": 200, "first_byte_us": 38900 } ]}Hopbits checks your site or API on a schedule, from home connections and datacenters in the countries you choose. Several nodes have to agree before it calls an outage, and it tells you whether one country fails or all of them.
Why
Most monitors look at your site from a few datacenters. That is not where your customers are.
A check from one place sees one route to your site. When a provider in another country cannot reach you, it stays green and you hear about it from a customer.
People reach you over home connections. Hopbits checks from those, and from datacenters, so you see both the user's road and a steady baseline.
One bad connection should never wake you. Hopbits asks several nodes on different networks, and calls an outage only when they agree.
How it works
Set a monitor up in the Console, through the API, or from a file in your repository.
Give it an address, how often to check, and up to three countries. Or let it use any node that is free.
On every run, up to three nodes in each country try to reach it, each on a different network. A country counts as down only when at least two fail and a second check from other nodes agrees.
Every change of state goes out by email and, if you want, as a signed webhook. When two runs in a row are fine again, the all-clear follows.
What you get
No feature is held back for a larger plan. A package is its credits and nothing else.
Pick one to three countries where your users are. Partial means one of them cannot reach you. Down means two or more of them, or all.
Checks run from real home connections and from datacenters, never from one operator alone.
From once an hour to every five minutes on any site. Every 60 or 30 seconds on a domain you have verified.
A light HTTP check works on any site. A full HTTP GET, a TCP connect and ping need a domain you have verified.
Show your visitors whether each service is up, with 90 days of history and your outages. One page for each website or project, at an address you choose.
Each change reaches your inbox and your own tools. The webhook is signed, so the tool that receives it can trust it.
Every monitor keeps the last 24 hours: the typical response time and the slow end, by country, with the latest answer from each node.
Uptime counts only the runs where enough nodes answered. A check that could not run never lowers your figure and never raises an alarm.
Keep your monitors in a file beside your code, see what a change will do, and apply it from your build with the Hopbits command line tool.
In practice
Four situations teams bring to it.
Your own office reaches the checkout without trouble, and the support inbox fills with complaints from one country.
The monitor turns Partial: one of your countries cannot reach it, the others can. You know it is a regional problem within one run, and where to look.
You build on a cloud, a CDN or a payment API, and its status page is still green while your errors climb.
Watch its public endpoint with a light check. When it fails you see it yourself, and one test from more countries shows how far the failure reaches.
You run an RPC endpoint or a gateway that people in many countries depend on every minute.
On a domain you have verified, check it over HTTP, TCP or ping as often as every 30 seconds, from the countries that matter to you.
An outage at night is found in the morning, from a customer's message.
Every change of state goes out at once, by email and as a signed webhook, so your own tools can open a ticket or page whoever is on call.
Tools
Everything the Console does is one call to the REST API. Teams who keep their setup in a repository write their monitors in a file and apply it from their build.
POST https://api.hopbits.io/v1/checks{ "name": "Public API", "type": "http", "target": "api.example.com", "method": "HEAD", "path": "/healthz", "interval_seconds": 300, "locations": ["DE", "US", "SG"]}201 Created{ "id": "chk_4b1d9e07a3c2f581", "name": "Public API", "type": "http", "target": "api.example.com", "method": "HEAD", "path": "/healthz", "interval_seconds": 300, "locations": ["DE", "US", "SG"], "alert_email": true, "verified": false, "paused": false}version: 1 monitors: - key: marketing-site name: Marketing Site type: http method: HEAD target: https://www.example.com interval: 5m locations: [DE, US, SG] alert_email: true# See what a change will do. Nothing is changed.$ hopbits plan -f hopbits.yaml # Make it so.$ hopbits apply -f hopbits.yamlMonitors in the docs Status pages Checks API Command line and monitors as code Verify a domain
Cost
Every account gets 1,000 free credits a day. That keeps one site checked every five minutes from one country, around the clock. More credits come by the month or as a one-time top-up.
The packages
Questions
More in the docsHow monitors work
Other productsNetwork testsIP location
Yes, with a light check every five minutes to every hour. Checks every minute or every 30 seconds, TCP and ping need a domain you have verified.
A country counts as down when at least two nodes on different networks cannot reach your site and a second check from other nodes agrees. A node that fails alone never counts against your site.
Not today. Any HTTP answer counts as up, an error page included. A monitor tells you whether your site can be reached and how fast it answers.
Through the webhook. Every change of state is sent, signed, to an address you choose, and your own tool turns it into a message or a page.
Yes. Switch on a status page for the monitors you pick: a public page with the state of each service, 90 days of history and your outages. You can have one for each website or project, and showing it uses no credits.
Checks wait until you have credits again. The free 1,000 come back every day at 00:00 UTC. A check that was skipped never raises a false alarm.
As many as you like. They all run on the same credits.
Get started
Create an account and you have 1,000 credits every day. No purchase. No wallet.