Features / Self-serve
Status page
One place to say what's wrong
A public page for when something is broken, so your inbox does not fill up with forty people telling you the same thing.
All systems
No status to report, because nothing is being checked yet.
Not connected to real monitoring, so nothing above has been measured. A status page showing uptime it never checked is worse than having no status page, so it stays off until the checks are real.
What it does
- Public page on your own domain
- Current state and the history behind it
- Linked from the widget and the portal
- Subscribers told when it changes
Without it
During an outage your support queue becomes the status page
Forty people tell you the same thing at once
The API is down. The queue fills with one message written forty ways, and the person who could actually fix it is busy answering them.
The only place to say 'we know' is everywhere
You post in Slack, reply to eight tickets, write a tweet, then start again ten minutes later with an update, because there is no single page to point people at.
Afterwards there is no record
A customer asks whether you were down last Tuesday, because their team lost an afternoon. The honest answer lives in a chat scrollback and somebody's memory.
How it works
Watch your own endpoints
HTTP and HTTPS checks against the URLs you choose, grouped by product, each with an expected status code, a timeout and a check interval.
What that includes
- HTTP
- grouped by product
- each with an expected status code
- a timeout
- a check interval
Decide what is public
Every monitor is public or private. The private ones stay internal for your team; the public ones are the only things that ever appear on the page.
Publish a page and link it where people already are
A branded page showing current state and the history behind it, linked from the support widget and the customer portal, so someone about to write in sees it first.
What that includes
- the history behind it
- linked from the support widget
- the customer portal
Tell the people who asked to be told
Subscribers confirm their own address, hear when the state changes, and can leave from any message they receive.
What it means for you
Your queue stops being a duplicate detector
The people who would have written in read the page instead, and the ones who still write in are told the same thing, once, without anyone retyping it.
You write the sentence once
One update goes to the page, the widget, the portal and the subscribers together. During an incident that is an hour of your afternoon back.
The history is on the record
When someone asks what happened last Tuesday, you send a link instead of an apology and a guess.
The honest limit: The status page is not live yet. The screens exist in the product, but they are not connected to real monitoring, so nothing on them has been measured. A status page showing uptime it never checked is worse than having no status page, so it stays switched off until the checks are real.