Alerts Timeline view in Azure

The beginning of autumn was relatively poor on interesting announcements in Azure, but I managed to catch something noteworthy last week. Let’s talk about Alerts — a mechanism for notifying about events that occur in the ecosystem.

Without going into too much detail, setting up an alert involves the following:

1. Define a Metric and a condition to fire. Human language may sound like “more than 50% CPU,” “number of errors exceeding five per minute,” etc.
2. Create an Action Group — a resource that defines notification methods and a list of recipients. Examples are emails, webhooks, phone numbers for SMS, or push notifications. But it could also include more sophisticated integrations like Logic Apps, Event Hubs, Azure Functions, Runbooks, etc.
3. Connect #1 and #2.

As the system grows, the number of configured alerts inevitably increases, so there is a need to interpret them effectively. I want to highlight “effectively” here because receiving tons of signals makes no sense unless you analyze the gathered information thoroughly. A forever-screaming alert is pointless because the only thing it does well is contribute to global warming.

Assuming we are motivated enough to investigate, we have been tackling an obstacle so far. There simply wasn’t a convenient tool for such…