Azure Monitor Alerts Pricing for Website Downtime Alerts

Azure Monitor alerts are cheap. A metric alert costs $0.10 per monitored time series a month after the first 10, a log search alert running every 5 minutes costs $1.50 a month, and notifications are nearly free: 1,000 emails, 100 US text messages and 10 voice calls a month at no charge, then $0.00645 per text and $0.013 per call. For downtime alerting, the expensive line is not the alert at all. It is the availability test that feeds it, at $0.0005 per execution, which is $21.90 a month for one URL at Microsoft's recommended setup.

Microsoft's pricing page draws its rates in the browser, so most summaries of it describe the categories and skip the numbers. Every figure below comes from Azure's public retail price list for East US in US dollars, and the limits come from Microsoft's Azure Monitor documentation.

How much do Azure Monitor alert rules cost?

It depends on the rule type. Metric alerts bill per time series they watch, and log search alerts bill per rule by how often they run. Activity log and service health alerts carry no rule charge.

A time series is one combination of metric and dimension values. An availability alert on a single test is one time series. Split the same rule by test location and five locations become five time series, which is the detail that turns a 20-test setup from $1 a month into $9.

What do Azure Monitor notifications cost?

Notifications go out through action groups, and each channel has its own free allowance before billing starts.

To put the SMS rate in perspective: a team of five on call, receiving a down and a recovery text for 40 incidents in a month, sends 400 texts. After the 100 free ones, that is $1.94. Voice calls for the same 40 incidents, one call per person when the alert fires, are 200 calls and $2.47.

What an uptime alert on Azure really costs

Here is a complete downtime alerting setup for 20 public URLs, priced three ways: Microsoft's recommended five locations every five minutes, the same with the alert split by location, and the cheapest configuration Azure allows.

Test executions are 43,800 a month per URL at the recommended setup (five locations, 8,760 runs each in a 730-hour month) and 2,920 at the thinnest (one location every 15 minutes). Log ingestion for the results is left out because it depends on your workspace; it is small next to the executions but not zero.

The pattern is the whole story. Alerting accounts for about 1% of the recommended bill. The other 99% is the cost of checking, and it cannot be sampled, capped or moved to a cheaper tier the way telemetry can. The full per-URL breakdown, with a calculator, is in our Azure Application Insights pricing breakdown.

The limit that matters more than the price

Azure Monitor rate-limits notifications per phone number. In a production action group, a number receives no more than one SMS every five minutes and no more than one voice call every five minutes, and the limit applies across all your subscriptions at once. An action group can also hold at most 10 SMS actions and 10 voice actions.

For downtime alerting this is a real design constraint. Outages rarely arrive one URL at a time: a shared database, a bad deploy or an expired certificate on a load balancer takes several endpoints down within the same minute. The first alert reaches the engineer's phone. The rest, for the next five minutes, do not arrive as texts. If your process depends on each down URL producing its own page, route alerts through a webhook into an incident management tool that pages the right engineer and escalates instead of relying on SMS alone. Microsoft applies these rate limits to SMS, voice, push and email only; webhooks are not rate-limited that way.

Is Azure Monitor cheaper than a dedicated uptime monitoring tool?

For alerting alone, Azure Monitor is close to free and hard to beat. For uptime monitoring as a whole, it depends almost entirely on how many URLs you check and from how many places.

Up to six URLs checked from one location every 15 minutes, Azure is cheaper, and if the only question is whether one internal health endpoint answers, keeping it in Azure makes sense. From the seventh URL at the thinnest setup, and from the very first at the recommended one, a flat per-monitor plan costs less. UptimeHub checks 20 endpoints every minute on Starter and 100 every 30 seconds on Pro, confirms each failure from a second region before alerting, and sends alerts by email, SMS, Slack and webhook with a branded status page included.

How do I reduce Azure Monitor alert costs?

Start with the rules, not the notifications. Use metric alerts where a metric exists, since the first 10 time series are free and each after that is $0.10, against $1.50 for a log search alert doing the same job every 5 minutes. Avoid splitting by dimensions you never act on. Run log alerts every 15 minutes where a 15-minute delay is acceptable, which cuts the rule price from $3.00 to $0.50. Delete log search rules you no longer need, because they bill on their schedule whether or not they ever fire.

If availability tests are the big line, which they usually are, the choices are fewer locations, longer intervals or fewer URLs, and each costs detection speed. Moving the public, customer-facing checks to an external monitor and keeping Azure alerts for internal metrics splits the job along the line where each pricing model is cheapest. Other per-run monitoring meters are compared in our synthetic monitoring pricing comparison, and every major vendor is lined up in the uptime monitoring pricing comparison.

Are Azure Monitor alerts free?

Some are. Activity log alerts, service health alerts and the first 10 metric time series carry no rule charge, and notification channels all start with a free monthly allowance. Log search alerts are never free: the cheapest runs every 15 minutes for $0.50 a month. In practice most small teams pay a few dollars a month for alerting and much more for whatever produces the signal.

Do Azure alerts notify you when the site comes back up?

Yes. Availability alerts send one notification when the alert fires and another when it resolves, and do not repeat while the site stays down. Microsoft's example is a site down for an hour with a 15-minute evaluation frequency: you get one email when it goes down and one when it recovers, not four reminders. That keeps notification counts, and costs, low, and it also means nobody is re-paged during a long outage unless your incident process does it.