Step 1
Detected
A check records a problem: downtime, an expiring certificate, a DNS change, a broken link. Uptime failures are verified from a second location first.
- Bitmovin
- HBO Nordic
- Obsidian
- Laravel
- Fathom Analytics
- PHP 8
- Stanford University
- Takeaway.com
- IGN
- VRT NWS
- spatie
Detection and delivery
The check finds it. The alert goes where you routed it.
The checks watch your sites from outside your stack. Notifications carry what they find to the destinations you set up. They are two separate jobs.
Step 2
Routed
Every destination that applies to the site, through the team, a tag, a tag group or the site itself, gets the alert if it accepts that event.
Step 3
Acknowledged
Acknowledge from the email, Slack or Telegram alert. Repeat alerts for that incident stop.
Step 4
Recovered
The recovery notification still comes through, and ends the acknowledgement. PagerDuty incidents and Opsgenie alerts are resolved for you.
Notifications go to your team, not to your client. Use a status page to keep clients informed.
Built for a client roster
One inbox for every client is noise.
An agency does not need one more alert. It needs each alert to reach the person who looks after that site. Add destinations at four places, and one client's outage never pages another client's team.
Team
Team destinations cover every site.
Set them up once and they apply to every monitor in the account. Most teams start here.
Example
One Slack channel and one mailbox get downtime alerts for the whole portfolio.
Site
Site destinations cover one monitor.
Add destinations that only apply to that site. A site can also stop the team, tag and tag group alerts for a channel, so only its own destinations on that channel fire.
Example
The client on a support contract gets its own PagerDuty rotation. The rest stay on the team channels.
Tag
Tag destinations cover every site with that tag.
Tag your sites, then attach destinations to a tag. A site with several tags gets the alerts of every tag it carries.
Example
Every site tagged wordpress also alerts the developer who maintains the plugins.
Tag group
Tag groups cover exactly the sites you mean.
A tag group only applies to sites that carry all of its tags. Wildcards match naming conventions: production* also matches production-eu and production-us.
Example
Only sites tagged both client and production get these alerts.
client
and
production
12 ways to hear it
Pick the channel your on-call person actually watches.
Turn on one channel or all of them. Every channel can carry every check, and each destination keeps its own settings.
Chat
- Slack
- Microsoft Teams
- Discord
- Google Chat
- Telegram
Alerts land in the rooms your team already has open, with a link to the full report. Slack and Telegram alerts carry buttons to acknowledge or check again.
Email, push and SMS
- SMS
- Pushover
- ntfy
Email is on from the start. SMS goes through your own Vonage account. Pushover and ntfy let you set a priority, and ntfy works with ntfy.sh or your own server.
On-call
- PagerDuty
- Opsgenie
New problems open an incident or alert in the rotation you already run. When the check recovers, we resolve the PagerDuty incident and close the Opsgenie alert.
Your own systems
- Webhooks
Every event you pick arrives as a signed JSON POST on an endpoint of yours. Log it, open a ticket, or feed your own dashboard.
Oh Dear provides us with real-time notifications about platform-wide issues ahead of customer support reporting. We can confidently tell our customers “we're on it”.
Sander de Vos, Laravel developer
The alert only helps if it reaches someone before the customer does.
Act without opening Oh Dear
Acknowledge it where you read it.
Acknowledge an alert from the email, Slack or Telegram message itself. Slack and Telegram can also run the check again, so you see straight away whether the fix worked.
#incidents
Illustrative
example.com seems down!
Error: HTTP 500. Downtime verified from Paris, France and Frankfurt, Germany.
Acknowledge alerts
Five minutes
30 minutes
An hour
Four hours
A day
A week
Tomorrow at 09:00
10:44
Jane acknowledged the uptime check for example.com for 1 hour. We'll mute repeat alerts and still let you know when it recovers.
11:02
Re-checked by Jane, this check has recovered.
- Acknowledge. Mutes the repeat alerts for this incident. It ends when the check recovers or the time you picked runs out, whichever comes first.
- Check again. Runs the check right away from Slack or Telegram, so you know whether the fix worked.
- Snooze. For planned work on a healthy check, set from the check page in Oh Dear. It mutes everything, recovery included, until the timer runs out.
- Everywhere else. Discord, Microsoft Teams, Google Chat, Pushover and ntfy alerts link to the full report in Oh Dear.
Want all the technical details?
Who gets what
Marketing gets broken links. DevOps gets downtime.
Each destination picks the events it receives and the hours it receives them. Match the alert to the person who owns it, instead of copying everyone.
Team destinations
Illustrative
Destination
Events
Hours
marketing@example.com
Broken links, mixed content
All hours
#devops
Uptime, certificate health
All hours
On-call rotation
Uptime
Outside business hours
hooks.example.com
Every event
All hours
Per destination
Business hours
Set your team's business hours. Then choose per destination: all hours, only inside business hours, or only outside them.
Per site
Maintenance periods
Plan a maintenance period and Oh Dear sends no notifications for that site while it runs.
Per destination
Failed deliveries
When a destination can't be reached for over a day, we pause it and email your team, so a dead webhook or Slack token doesn't fail quietly.
No feature tiers
Every check.
Every plan.
Notifications come with every Oh Dear check on every plan. Choose by the number of sites you monitor, not by which channels you need.
Start with everything on.
Try every channel alongside uptime, DNS, certificates and the rest of the website-health checks.
- Every channel on every plan
- Unlimited notification destinations
- Every monitoring feature included
- Plans scale by monitored site count
10-day free trial. No credit card.
Before the first alert
Questions about notifications.
What notifications carry, where they go, and where this page ends.
What are Oh Dear notifications?
They deliver what the checks find. Downtime, certificate problems, DNS changes, broken links and every other check result can raise an alert on the channels you set up. Notifications do not check anything themselves; they carry the findings to the right people.
Which notification channels does Oh Dear support?
12 channels: email, Slack, Microsoft Teams, Discord, Google Chat, Telegram, SMS, Pushover, ntfy, PagerDuty, Opsgenie and webhooks. Every new account starts with an email notification to the address you signed up with. You add the others yourself.
How fast will I be notified if something is wrong?
As soon as a check records the problem. For uptime, a failure seen from the primary location is verified from a second location first, so a single network hiccup does not wake anyone up. Then every destination that accepts that event gets the alert.
Can I send alerts to more than one place at once?
Yes. Add as many notification destinations as you like and mix them freely: a single check can ping your team's Slack, email a client, and page whoever's on call through Opsgenie, all at the same time.
How do I route alerts for different clients?
Add destinations at four places: the team, a single site, a tag, or a tag group. A site gets the alerts of every place that applies to it. A tag group only applies to sites that carry all of its tags, and wildcards like production* match your naming conventions. Read the routing guide.
Can different people get different alerts?
Yes. Each destination has its own list of events. Marketing can take broken-link notices while DevOps takes downtime and certificate problems, in the same account.
Can I act on an alert without opening Oh Dear?
Yes. Email alerts have links to acknowledge them, and Slack and Telegram alerts have buttons to acknowledge them or run the check again. An acknowledgement mutes the repeat alerts for that incident and ends when the check recovers or the time you picked runs out. You still get the recovery notification.
Can I keep alerts quiet at night or during maintenance?
Yes. Set your team's business hours, then choose per destination whether it gets alerts at all hours, only inside business hours, or only outside them. During a maintenance period, Oh Dear sends no notifications for that site.
What if the person on call has no data connection?
Use SMS. Text messages travel over the mobile network, so they arrive without mobile data or Wi-Fi. Oh Dear sends them through your own Vonage account, so you add your Vonage API key and secret first.
How do webhooks work?
Oh Dear sends a JSON POST to your endpoint for each event the destination accepts. Every request is signed with your team's secret in the OhDear-Signature header. If your endpoint does not answer with HTTP 200, we try again, up to three times in total. See the webhook payloads.
Do notifications tell my client?
Not by default. Notifications go to the destinations you set up, which are usually your own team. To keep clients informed during an incident, use a public status page.
Is there a limit to how many notifications I can set up?
No. Notification destinations are unlimited on every Oh Dear plan, so you never ration channels or pay extra to add another. The only thing that scales with your plan is the number of sites you monitor.
Get started
Hear about it before the client does.
Add your sites and send the alerts to the channel your on-call person watches. Every feature on. No credit card.
