Everything here works today

Each item below is live and available from the console or the API. If something is not on this page, Statuskeel does not do it yet.

What your visitors read.

Address: acme.statuskeel.com, or a subdomain of yours such as status.acme.io with one CNAME record. The certificate is issued for you; we never ask for access to your DNS.
Components: Grouped, described and shown in the order you choose.
Statuses: Operational
Under maintenance
Degraded performance
Partial outage
Major outage
Unknown
Uptime: A bar for each day, up to 90 days depending on your plan. A major outage counts in full, a partial outage half, degraded performance a quarter, and maintenance windows are left out.
History: Correct the history by hand. The edit is marked as manual, kept in the activity log, and the uptime is counted again.

What your team writes.

Lifecycle: Investigating
Identified
Monitoring
Resolved
Impact: No impact
Minor
Major
Critical
Set per component. The components follow it for as long as the incident is open.
Templates: Save the incidents you write again and again, and start the next one from a template.
Postmortems: Add a postmortem to an incident, and it appears on the incident’s public page.
Maintenance: Scheduled windows start and finish themselves, and subscribers are warned before they begin.
Console: In English, Russian and Serbian, with an activity log of who changed what.

What runs while nobody is looking.

HTTP checks: HTTP and HTTPS: the expected code, a latency limit, a timeout, and an interval from 30 seconds to an hour, depending on your plan.
Agreement: A status changes only when at least two independent checkers agree: after three bad answers in a row by default, and back after two good ones.
Silence: When the checks stop reporting, the page says so and shows Unknown rather than green.
Incidents: A failing check opens its own incident, closes it on recovery, and reopens it if the outage returns within 30 minutes.
People first: An incident your team opened holds the component for as long as it is open, and a status set by hand holds it for an hour; the checks wait.
Metrics: Charts from Prometheus, Datadog, or numbers you push through the API, shown next to the components.

Who hears about it, and how.

Channels: Email with double opt-in, Slack, and webhooks signed with HMAC-SHA256.
Scope: Subscribe to the whole page, or only to the components you depend on.
Feeds: RSS and Atom feeds of incidents, for readers and tools that poll.
Delivery: Five attempts with a growing pause between them, and one message per update, never two.

From one record to a page that watches itself

Four steps. After the first two, Statuskeel does the rest.

connect

Create the page and point a subdomain of yours at it with one CNAME record. The certificate is issued once the record resolves.

; at your DNS provider
status.acme.io.  CNAME  acme.statuskeel.com.

✓ record verified
✓ certificate issued over HTTP-01

watch

Add an HTTP check to a component: the URL, the expected code, a latency limit and how often to look.

url       https://hooks.acme.io/health
expect    200
slow      after 800 ms
timeout   5 000 ms
every     60 s
fail      3 in a row, recover after 2

No plugins to install: everything below works through the REST API, incoming webhooks and outgoing notifications.

Status from: Alerts and scripts that change a component’s status.
Prometheus Alertmanager
generic JSON webhook
REST API
Metrics from: Charts shown next to your components.
Prometheus
Datadog
your own numbers through the API
Notifications to: Where subscribers hear about incidents and maintenance.
email
Slack
signed webhooks
Read as: For people and tools that poll.
RSS
Atom
status JSON
SVG badge

Create the page in the console, add your components, and point a subdomain of yours at it when you are ready.

DNS · acme.ioone line

; at your DNS provider
status.acme.io.  CNAME  acme.statuskeel.com.

✓ record verified
✓ certificate issued, renews itself
$