Services checks allow you to characterize the status of a service to monitor it within Datadog. Service checks monitor the up or down status of the specific service. You are alerted whenever the monitoring Agent fails to connect to that service in a specified number of consecutive checks. For example, you can get an alert any time the monitoring Agent on a Redis host reports three consecutive failed attempts to connect to Redis and collect metrics.
Service checks at the cluster level offer another effective way to monitor distributed or redundant systems that can withstand some failures. Use these alerts for architectures where individual hosts run multiple services, because they can surface the degradation of the service even if the hosts running that service remain available (and would pass a host-level health check).
You can set up monitoring and an alert for when a critical, non-redundant service is lost, or if a cluster is on the verge of failure due to widespread node loss. Other critical alerts could be a drop in request throughput or an increase in request latency.
You might need to set up a service check if an integration does not have one natively or for an internal service that you want to monitor for up or down status.
To use service checks, first set up the check:
Additional helpful documentation, links, and articles:
Once the service check is sending data, check out your check summary and set up dashboards, monitors, and alerts:
Service checks can be visualized and used in 3 Datadog sections:
Click on the Monitors tab, and then on Check Summary to see the Check Summary page.
This is a list of all your checks and their statuses, reporting across your infrastructure in the past day. Select a check to get insights on the number of distinct checks for each of its associated tags.
You can visualize service checks with a Check status widget in a screenboard:
After clicking on the Check status widget icon, the following pop-up appears:
In this form, you can:
Even if you can’t graph a custom check over time as you would for metrics, you can still monitor it. Go to monitor tab > new monitor, then select the custom check section:
Configure your custom check monitor:
In this form, you can:
Service check monitors for out-of-the-box Datadog integrations can be set in the monitor tab > Integration > Integration status tab on the left. For instance, with the HAProxy integration:
Learn how to write an integration with the integration developer documentation.