Before we continue with alerting, let's review the other options on the scenario level:
- Attempts: Web pages are funny beasts. They mostly work, but that one time when the monitoring system checks it, it fails. Or is it just that users reload a page that fails to load once and never complain? No matter what, this field allows us to specify how many times Zabbix tries to download a web page. For pages that experience the occasional hiccup, a value of 2 or 3 could be appropriate.
- Agent: When a web browser connects to a web server, it usually sends along a string identifying itself. This string includes the browser name, version, operating system, and, often, other information. This information is used for purposes such as gathering statistics, making a specific portion of a site work better in some browser, denying access, or limiting experience on the site. Zabbix web monitoring checks also send user-agent strings to web servers. By default, it identifies as Zabbix, but you can also choose from a list of predefined browser strings or enter a custom string by choosing the Others option:
- HTTP proxy: If needed, an HTTP proxy can be set per scenario. A username, password, and port can be specified as well:
The default HTTP proxy can be set with the http_proxy and https_proxy environment variables for the Zabbix server process – these variables will be picked up by libcurl, which is used underneath for the web monitoring. If a proxy is specified on the scenario level, it overrides such a default proxy setting. There is no way to set a proxy on the step level.
We'll discuss the remaining fields, Variables and Headers, a bit later in logging into the Zabbix interface.
Web monitoring in Zabbix does not support JavaScript at all.