01 / First contact
Up and watching.
Holler checks whether your websites and endpoints respond as expected. If a site goes down, we send an email. When it comes back, we holler again.
- Sign in with your email.
We’ll send a six-digit code. New here? The same flow creates your account—there’s no password to remember.
- Add your first monitor.
Give it a name you’ll recognize, paste its public URL, and choose how often to check it.
- Let the observations arrive.
The dashboard will show the result once the first check completes. A new monitor starts as unknown, not automatically healthy.
02 / Your watchlist
Make each check count.
Open Overview → Add monitor. Holler supports public HTTP and HTTPS endpoints on ports 80 and 443, including redirects. Local devices and private network addresses aren’t supported.
| Interval | A useful starting point | Confirmed down, approximately |
|---|---|---|
| 1 minute | Your main site or API | 1–2 minutes |
| 5 minutes | Less time-sensitive services | 5–10 minutes |
| 20 minutes | Occasional availability checks | 20–40 minutes |
Two failed checks confirm an outage. Request time and scheduler delays can add to these times.
The default expected response is HTTP 200, with a 10-second timeout. Responses above 1,500 ms are marked slow by default. You can adjust these settings when creating a monitor.
Taking something offline on purpose?
Pause stops checks. Maintenance keeps checking but excludes those observations from availability and suppresses incident transitions. An existing incident resolves after a successful check outside maintenance.
Check now queues an extra observation; it does not mean the check has already succeeded. To change a target URL or expected status, create a new monitor so its history stays accurate.
03 / When we holler
A heads-up. And an all-clear.
One failed check marks a service as degraded. A second consecutive failure opens an incident and queues a down email. One successful check resolves the incident and queues a recovery email.
Failed check → degraded
Second failure → incident + down email
Successful check → resolved + recovery email
Mail comes from [email protected]. In Team, each signed-in person can choose whether to receive alerts for that workspace. Pending invitees don’t receive alerts.
An email accepted by our sending provider isn’t proof of inbox delivery. Check spam filters if an expected alert doesn’t arrive. Uncertain sends are held rather than blindly resent.
04 / Keep people in the loop
Your status, shared simply.
Choose which monitors are public, then use Public status page in the header to open your shareable page. Workspace settings let you change its name and URL slug.
Open Status pages to create additional pages and select a different set of public monitors for each audience. Each extra page needs a $0.99 monthly page add-on. The included page always shows all public monitors.
The page shows shared services, recent observations, availability history, and incident updates. Private monitors and target URLs are omitted. Public service names and incident notes are visible to anyone with the link, so keep confidential details out of them.
You can post an incident update to explain what happened or what your team is doing. Readers can also use Read this page as JSON for the same public observations.
05 / Shared responsibility
Bring the people who keep things running.
Open Team → Invite a team member and enter their email. They sign in with that exact address, review the invitation, and choose whether to accept alerts.
- Owners manage monitors, settings, API keys, and team membership.
- Members manage monitors and incident updates.
- Invitations expire after seven days. Sending one grants no access until it’s accepted.
The workspace selector switches between your own workspace and teams you’ve joined. Removing a member ends their team access and future alerts without deleting their personal workspace. Up to 20 members and pending invitations can belong to a workspace.
06 / What the numbers mean
Observed. Never assumed.
Availability
Successful checks divided by eligible checks. Maintenance is excluded; gaps aren’t invented successes. “Check now” observations are included.
Response time
Time from a GET request to response headers, including redirects. It’s not full page-load speed or a browser rendering test.
Checking location
The named server making the observation. This version checks from one location, not from every region or every visitor’s connection.
Unknown & gray history
No fresh observation or no data for that period. After two intervals plus 30 seconds without a result, a monitor becomes unknown.
Availability windows cover 24 hours, 30 days, and 90 days. Raw checks are retained for 90 days; recent-check history returns the latest 200. Incident lists show up to 100 incidents, with open incidents first.
07 / Humans & machines
Same signal. Your interface.
Prefer a script or agent? Open Agent access to create a scoped key. Keys are shown once, stored hashed, and stop working when revoked.
- Read: inspect workspace observations.
- Write: also manage monitors and incident updates.
- Admin: also manage keys and workspace settings.
REST API · create a monitor
Set HOLLER_API_KEY securely in your environment, then:
curl https://hollerifitsdown.com/api/v1/monitors \
-H "Authorization: Bearer $HOLLER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"name":"My website","url":"https://example.com","public":true}'PUT replaces editable settings; preserve values you want to keep. A check request returns 202 when queued. After an uncertain create request, inspect the list before retrying.
Python CLI · work from your terminal
From a Holler source checkout, using Python 3.9 or newer:
export HOLLER_URL="https://hollerifitsdown.com" # Set HOLLER_API_KEY securely in your environment. python3 tools/holler.py monitors python3 tools/holler.py add "My website" https://example.com python3 tools/holler.py status your-page-slug
MCP · connect an agent
Run python3 tools/mcp_server.py as a stdio MCP server with HOLLER_URL and a scoped HOLLER_API_KEY in its environment. The agent uses the same API and permissions as the dashboard. This version has no hosted HTTP MCP endpoint.
Status notes and incident text are author-controlled content, not instructions for your agent to follow.
08 / Back on course
A few things worth checking.
My sign-in code hasn’t arrived.
Check your address and spam folder. Codes expire after ten minutes and allow five attempts. Wait at least a minute before requesting another; repeated requests are rate-limited.
My browser loads the site, but Holler says it’s down.
Holler checks from its server, so your firewall, bot protection, location rules, or expected HTTP response may differ from your browser’s experience. Look at the monitor’s latest HTTP status and error details. A login page or redirect can return a different result than you expect.
I don’t see my teammate’s workspace.
Sign in with the exact invited email address and accept the invitation. Then choose the team in the workspace selector. Ask the owner for a new invitation if the old one expired.
Where are my alerts?
Confirm that your workspace alert preference is enabled and the service failed twice outside maintenance. Check the Notifications API for provider outcomes. An accepted email can still be delayed or filtered by your mailbox.
09 / Room to grow
A little extra. For a little extra.
Every workspace includes 10 monitors and one public status page, free. Team membership is included within the workspace’s 20-person limit.
| Add-on | Capacity per pack | Monthly price |
|---|---|---|
| Monitor pack | 10 additional monitors | $0.99 |
| Status page | One additional public page | $0.99 |
For example, two monitor packs and one page add-on cost $2.97/month and give your workspace 30 monitors and two public pages. You can choose up to 100 of each add-on.
Buying and changing packs
Sign in as the workspace owner and open Billing. Choose quantities, review the monthly total, then confirm. Your first purchase opens Stripe Checkout. An unfinished checkout resumes the same selection shown on Stripe; you can change quantities after completing it or after that checkout expires.
Changes to an existing subscription charge or credit the prorated difference immediately. Extra capacity activates only after payment is confirmed. If Stripe needs another payment step, finish it on the invoice page. Pending or failed upgrades do not grant the requested extra capacity.
Invoices, payment methods, and cancellation
Manage billing opens Stripe’s secure portal for invoices, payment details, and cancellation. Canceling there keeps your paid capacity until the end of its paid period and stops renewal. Reducing pack quantities applies immediately.
Admin API keys can manage pages and read capacity. Purchasing and managing payments requires the signed-in owner’s identity; a newly created admin key cannot spend money. The installation owner may use their private owner key.
Billing support: [email protected]. Card information is entered on Stripe; Holler stores customer and subscription identifiers, not card numbers.