Skip to content

Mattermost alerts from outside your firewall.

Uptimia checks your sites from 171+ locations in 70+ countries and posts what breaks into Mattermost through one incoming webhook. Nothing is installed on your server — no plugin, no bot account, no token that could read your channels.

Green graphs, unreachable site

Self-hosting means you already run monitoring — Prometheus, Zabbix, node exporters — all watching from inside. Failures customers hit — expired certificates, DNS, the edge — are outside, where collectors never look.

Inside the network

Every graph says healthy

Cert expires
07:58 · at the edge
invisible inside
Grafana: all green
services up · disks fine

The app is running, the containers are happy, and the LAN doesn’t do TLS to itself. From inside the network, an expired certificate simply doesn’t exist.

Outside, meanwhile

Browsers are showing red screens

Cert expires
07:58 · public
every visitor
Your connection is not private
bounce · bounce · bounce

Customers don’t see your graphs — they see the browser warning and leave. The gap between those two views is exactly where outside-in monitoring lives.

Internal monitoring proves the services are up. Uptimia proves customers can actually reach them — and posts the difference to Mattermost. The same 07:58 expiry, wired up:

The certificate nobody renewed

At 07:58 the portal’s certificate expired. Grafana stayed green, every container stayed up, and every visitor got a browser warning — here are the 24 minutes that followed, seen from ~alerts.

07:58 Portal fails TLS from every regionCertificate expired — 3 outside regions confirm, ~alerts gets the post ~alerts
08:00 Jonas acknowledges straight from the channelSigned link, no Uptimia login — the ladder pauses on the spot ~alerts
08:07 The fix is worked out under the alert“cert-manager renewal hung since Friday — re-issuing now ~alerts
08:22 Recovery posts to ~alertsHandshake green from all regions — total time down: 24 minutes ~alerts
24 mindown → up
Your infra, your chat, your data0 new tools
The whole incident lived on systems you run — except the watching, which has to happen from outside. That’s the entire division of labour.
backup checks agreedacknowledged in 2 mindown 24 min1 webhook, total
No SMS storm, no phone tree. The acknowledgement at 08:00 paused the ladder before it ever left the channel. ladder paused

One incoming webhook, no plugin

Mattermost’s built-in incoming webhooks are the entire mechanism. Uptimia holds a URL it can only write to, and the destination channel is set inside Mattermost, not by us.

Step 12 min

Create the incoming webhook

In Mattermost: Integrations → Incoming Webhooks → Add. Pick the channel, copy the URL, paste into Uptimia and send a test. If it’s hidden, a system admin enables webhooks in the System Console.

~alerts Send test
test post visible in ~alerts · wiring proven
Step 21 min

Route monitors to channels

Add one webhook per channel and pick the destination per monitor: outages to ~alerts, certificate and domain warnings to ~ops-warnings. Because the channel lives in the webhook, moving a feed is a change you make in Mattermost.

Alert this monitor to
~alerts~ops-warningsEmail
Portal — caldmont.com → posts to ~alerts
Step 3optional

Set the failure mode

Decide what happens when nobody acknowledges: escalate to SMS, then to the next responder. Quiet hours and maintenance windows are honoured — a planned reboot pages no one. Escalation policies come with the Professional plan and above; the webhook itself works on every plan.

If nobody acknowledges — escalate
~alerts → +5 min SMS on-call → +15 min ~incidents
maintenance windows muted · quiet hours per person
Technical setup guide

Connect Mattermost to Uptimia

Screen by screen, in the Help Center: turning on incoming webhooks in the System Console, adding the integration, and keeping the webhook URL working afterwards.

  • Enable incoming webhooks in Mattermost
  • Add the integration in Uptimia
  • Test the integration
  • Attach it to your monitors
  • Keep your webhook URL working
Open the setup guide help.uptimia.com

Twelve monitor types, one channel

Node metrics won’t tell you the domain lapsed, the certificate chain broke, or that checkout dies on step three. All twelve types post through the same webhook — the eight below, plus DNS, blacklist, API chains and real-user data.

Uptime checksReachability from 171+ points
SSL certificatesChain & expiry, outside-in
Domain expiryRegistrar lapse dates
Server metricsAgent metrics, if you want them
1 mattermost channel ~alerts — everything posts here
Heartbeats & cronA dead-man switch for cron
TransactionsMulti-step flows in real Chrome
Virus & malwareBlacklist & malware flags
Page speedReal-browser load times
One attachment format across all twelve — red bar for a failure, green for a recovery, so the channel scans in one pass. 12 types · 1 channel

Three channels, three urgencies

Alert channels die of mixed urgency. Split the feed instead: only real outages in ~alerts, expiry warnings in a daylight channel, and performance numbers where they inform without interrupting.

~alerts
Wired to whoever is on call.
Uptime & HTTP checks
Checkout & transaction flows
Server metrics & heartbeats
urgent · allowed to page
~ops-warnings
This week’s work, not this minute’s.
SSL certificate expiry
Domain expiry
Virus & malware flags
scheduled · daylight only
~perf-watch
Trend lines for the next retro.
Page speed & Core Web Vitals
Uptime on marketing pages
Campaign landing pages
informational
Some alerts need to leave chat entirely? They also reach SlackMS TeamsDiscordTelegramPagerDutyEmail & SMS

Watch from the outside.Alert on the inside — in your Mattermost.

Outside-in checks, alerts on your server — the full platform, 30 days free.

Start your free 30-day trial
30 days free no credit card cancel anytime

Behind every Mattermost post

Nothing runs on your Mattermost server and nothing new joins it. Uptimia does the watching on its own fleet and hands the results to your channels.

The dashboard when you want depth

Incidents, response times, per-region results — the drill-down behind every channel post.

Portal — caldmont.comUPTIME · 99.97% Nightly backupHEARTBEAT · OK caldmont.comSSL · RENEWED

171+ vantage points

Checks run from 70+ countries, and you decide how many other regions must repeat a failure before the post goes out.

171+ checkpoints · 70+ countries

Status pages included

A public status page on your own domain — for the users who ask before you’ve posted.

 status.caldmont.com · live

Uptime reports

Monthly PDFs with the numbers — straight into the SLA folder.

June · 99.97% · PDF sent

Response times on record

Acknowledgements are timestamped, so MTTA is measurable — not anecdotal.

avg. response · 2 min

When the channel isn’t enough

On an escalation policy (Professional plan and above), unacked alerts climb: SMS to the on-call, then the next person. Maintenance windows keep planned work off the pager.

Step 1 — ~alertsMATTERMOST Step 2 — on-callSMS · +5 MIN Maintenance windowMUTED

How Mattermost website-monitoring alerts work

Uptimia monitors your sites from 171+ external locations and delivers alerts to Mattermost through a standard incoming webhook — the same mechanism whether your server is self-hosted or on Mattermost Cloud. Each post is a message attachment: a coloured bar, the monitor and what failed, Priority and Type fields, and — on monitors running an escalation policy (Professional plan and above) — a signed Acknowledge incident link. Recoveries post to the same channel with the total time down. The webhook URL is your server’s /hooks/ endpoint, which has to answer from a public address.

“We’ll add a Grafana alert”

Internal alerting stops at the firewall

Edge failure
cert · DNS · routing
no probe out there
No datapoint
nothing to alert on

You can’t graph what you don’t measure — and internal collectors never make requests from where your customers are standing.

Outside-in + Mattermost

External probes, internal channel

171+ locations
checking constantly
one webhook
~alerts
on your server

The measurement happens outside; the conversation stays on infrastructure you own. One HTTPS POST is the entire integration surface.

Every event

Exactly what arrives in your channel

Five posts cover the incident lifecycle — the channel ends up scrolling like an ops log.

See all 12 alert channels
PostArrives whenContents
Down alertBackup checks confirm the failureMonitor, cause, priority, type — plus the acknowledgement link on an escalation policy
Grouped digestMonitors on one policy fail togetherOne post listing every monitor that is down
Join updateMore monitors fall into the same incidentA throttled follow-up post with the new count
Escalation stepThe alert sits unacknowledgedClimbs to SMS or another channel, then the next person
RecoveryThe check passes againGreen bar, all-clear and total time down
Test postYou press Test on the integrationProof the wiring works, before anything real

Mattermost alerts FAQ

01Does Uptimia work with self-hosted Mattermost?+
Yes — that’s the primary case. Uptimia talks to the incoming-webhook endpoint every Mattermost server ships with, so there is no version to match and nothing to keep updated. Mattermost Cloud works the same way.
02Do I need to install a plugin or create a bot account?+
No. The integration is a standard incoming webhook — nothing is installed on your Mattermost server, no plugin, no bot user, no extra login. You create the webhook, Uptimia posts to it.
03What does Uptimia need to reach my server?+
One webhook URL that answers from a public address. URLs resolving to private, LAN or link-local addresses are refused when you save the integration, and an https endpoint has to present a certificate that verifies — an internal CA or self-signed cert will not deliver. If your server stays private, pair Uptimia with email or SMS instead.
04What data leaves my infrastructure?+
None — the data flows toward you. Uptimia POSTs alert content (monitor name, cause, timing, links) to your webhook; it never reads your chat, your users or anything else on the server. You share exactly one URL.
05What can it catch that our internal monitoring can’t?+
Everything that only exists outside your network: expired or broken certificate chains, DNS problems, domain expiry, edge and routing failures, per-region outages, and multi-step flows like sign-in or checkout replayed in a real browser.
06How does acknowledging work from Mattermost?+
Monitors running an escalation policy (Professional plan and above) get a signed Acknowledge incident link in the post — one click, no Uptimia login. It pauses the ladder rather than closing the incident, and records who responded and how fast; later posts about that incident carry the “Acknowledged by” line.
07Can different monitors post to different channels?+
Yes — create one incoming webhook per channel and each monitor picks its destination: outages to ~alerts, expiry warnings to ~ops-warnings, performance numbers to ~perf-watch.
08What if nobody sees the post?+
On an escalation policy — Professional plan and above — the ladder takes over: after a delay you choose, Uptimia sends an SMS to the on-call person, then moves to the next responder. An acknowledgement — from the channel link or the dashboard — pauses the climb.
09How noisy will the channel be?+
Quiet by design: uptime, SSL, speed and transaction failures are re-checked from other regions before anything posts — up to three have to agree, and you set the bar. Monitors on the same escalation policy collapse into one incident, maintenance windows silence planned work, and quiet hours hold the non-urgent posts.
10Is Mattermost support an add-on?+
No — Mattermost is one of 12 built-in alert channels, on every plan including the 30-day free trial. No credit card required.

Add the outside view

Create one incoming webhook, send a test post, assign your monitors — from then on, the internet’s view of your site reports to your channels.

Self-hosted or Cloud No plugin, no bot 30-day free trial No credit card
Mattermost is one of 12 built-in alert channels — every Uptimia monitor type can post to it.