v1.4, August 22, 2012

These release notes correspond to the "Unlimited Availability" release of Cloud Monitoring.

New Features

  • More HTTP Options: Pass in custom headers for a remote.http check. Depending on the site you're testing, this can be crucial in determining if your site is up.

    remote.http Docs

  • Add a previous operator to the Alarm Language: This allows you to look at a metric value in the previous time period. It is very useful for detecting changes. For example, to detect an SSH fingerprint change, you'd construct an Alarm snippet like this:

    if (metrics['fingerprint'] != previous(metrics['fingerprint'])) {
      return new AlarmStatus(CRITICAL, 'SSH fingerprint changed to #{fingerprint}');
    }

    Alarm DSL Docs

  • Improved email alerts: Included in emails are the observations of the particular check from each monitoring zone. This allows you to get visibility on the status from any of the included monitoring zones of a check. We've also reformatted the message to look more user friendly.

    Subject:

    ** CRITICAL: host server check on web1.domain.com **

    Body:

    ================== Rackspace Cloud Monitoring Notification ===================
    Entity: web1.domain.com (50.56.179.42)
    Check: host server check (remote.ping)
    Alarm: Ping packet loss
    Status: CRITICAL
    Date: Mon Aug 13 2012 15:18:59 GMT+0000 (UTC)
    
    --------------------------- Observations -------------------------------------
    
    CRITICAL - Packet loss is 20%
    London (LON), 25 sec ago
    
    OK - Packet loss is normal
    London (LON), 10 sec ago
    
    OK - Packet loss is normal
    London (LON), 56 sec ago
    
    ------------------------------------------------------------------------------
    
    This is an automated Cloud Monitoring notification. You received this email
    because you are listed as a notification recipient.
    				

  • Retrieve overview by entityId or uri: Include query strings uri or entityId to retrieve specific mapped Rackspace Cloud Servers (for now). You can also pull down a set of entities by referencing their id's. Keep in mind you cannot combine pulling both uri's and entityId's in the same request, but you can have up to 100 of them in a single request.

    Overview list by entityId and Overview list by uri.

Enhancements

  • Added friendly names for the alarm_examples API.

  • Add a "metadata" attribute to the alarm API.

  • Allow including no conditionals in the alarm language, allowing you to set global filters, and rely on the check availability to determine the returned state.

Resolved Issues

  • Fix un-setting optional attributes on an update (PUT).

  • Fix the check_type -> alarm mapping for delivering inconsistent and non-deterministic updates.

  • Fixed how we handle zlib streams when inflating a gzip http stream.

  • When we combined two services we had briefly did not insert the initial OK state for the alarm change log, this is fixed now.

  • Fixed updating the target_alias after setting the target_hostname on the initial creation, and vice-versa.