Search answers on uptime checks, webhook alerts, heartbeats, on-call rotations, incidents and the API — or ask the team.
Browse by topic
Our health endpoint returns 200 with a small JSON body. I can curl it from anywhere and it is fine. MonoDuty keeps flipping the monitor to DOWN and opening incidents. The error text mentions something about expected body.
We moved our cron jobs into Kubernetes and now some heartbeat check-ins fail with 429. Each job pings its own heartbeat and none of them run more than once a minute individually, so I do not understand why we are being rate limited.
We had Slack alerts working fine. Last week we archived the old Slack channel and made a new one, and generated a new incoming webhook URL. I updated the URL in MonoDuty but nothing arrives anymore. The MonoDuty channel is still listed as active.
I am trying to script monitor creation. I looked for an API key section in settings and could not find one, so I tried using my account password with basic auth and got 401. What is the correct way to authenticate against the API?
I installed the mobile app and signed in, I can see all our incidents in the app. But I never get a push notification when a new incident opens, even though my colleague on the same team does. I have checked that I am in the right organization.
I need to report monthly availability to my manager. I can see the individual check results in the dashboard but I cannot find a single uptime number anywhere. Is there a report or do I have to compute it myself from the raw results?
I want to build an internal dashboard showing only incidents that are still live, meaning open, acknowledged or investigating. Filtering by a single status works but I cannot figure out how to pass several at once without making three separate calls.
I added a notification channel of type webhook and put our internal endpoint URL in the config. The channel shows up in the list and is active. But when one of our monitors went down last night we got the incident in the dashboard and nothing arrived at our endpoint. The endpoint works fine when I curl it manually.
When I get paged at night I would rather not open a laptop just to acknowledge. Can I acknowledge and resolve incidents directly from the mobile app, or is it read only?
Our nightly job pinged the check-in URL on time, I can see the request in our own logs. MonoDuty still marked the heartbeat as down and opened an incident at the same minute the ping arrived. The incident description mentions something about the body not matching.
I invited a new engineer two weeks ago. He finally clicked the link today and gets a message saying the invitation is invalid or expired. I can still see the invitation in our pending list so I assumed it was fine.
I set up an escalation policy with three levels, email first, then SMS after five minutes, then a call after ten. In practice we only ever get the first email. Even when nobody responds for half an hour, the SMS never comes.
Our whole company runs on Microsoft Teams, we do not use Slack at all. I see Slack and Discord as channel options but nothing for Teams. Is Teams supported in some way I am missing?
Someone deleted an incident by mistake. I found it in the deleted list and restored it, which worked, the incident is back. But the timeline is empty and the people who were assigned to it are no longer there. Is there a way to get that back?
Our support team and a few managers only need to look at incidents, they never configure anything. I want to give them access without inflating our bill. Is there a cheaper or free level of access for people who only read?
We fire an alert trigger from our deploy pipeline and pass severity in the JSON body. No matter what we send, the incident always ends up with the severity we configured on the trigger itself. We send lowercase values like critical and warning.
We use Grafana alerting and want the alerts to open incidents in MonoDuty. I set up a contact point of type webhook pointing at an alert trigger URL. Incidents are created but every single one has the same generic title, none of the Grafana context comes through.
We are about to publish a status page for our customers. I am not sure whether to list every monitor we have or just a few. We have about forty monitors, many of them internal things like queue workers and database ports.
While testing our integration I accidentally pasted a truncated token and the request still returned OK. I then tried a completely random token and it also returned OK. That makes me nervous, how do I know our real check-ins are actually being recorded?
We just set up our first rotation and the on-call person is getting woken up by things that could easily wait until morning. I do not want to remove the alerts entirely because they are useful during the day. How do other teams handle this?
