Introducing the new, improved Apache Kafka specific alarms

Written by Magnus Landerblom

Users who may have been unsure of notifications and alarm settings can now rest easy. These new improvements will ensure the notification of cluster setbacks and other concerns. Cloudkarafka has made significant improvements to the configurable Apache Kafka alarms and addressed some popular user requests. Read on to find out about new improvements.

All clusters on Cloudkarafka are monitored to ensure that they are online and continue serving messages. The server is monitored, for the status of CPU, memory, and disk usage and our customers have the ability to set up alarms for any of these metrics, and now this service is even more robust!

Get notified by all triggered alarms

First, the system that notifies customers about the alarms has been updated. Previously, every alarm had a cooldown of one hour after it got resolved, meaning that if the same alarm would trigger again, there would be no notification. This is no longer the case, and users are notified about all triggered alarms.

Add a reminder interval

Now, a reminder interval can be specified for the alarms. The previous setting notified users once the alarm was triggered and also once it got resolved. With the upgrade, users have the ability to add a reminder interval so that if the alarm cannot be handled directly, there will be another notification after a specified interval. note: These intervals don’t apply for those using PagerDuty, OpsGenie, or VictorOps as they have their own way of handling reminders.

Image description: Flow chart of how the alarm system works

Great improvements for UI

Major improvements have been made to the UI parts. Existing alarms can be edited, including changing the time threshold, value threshold, and basically tweaks for all parameters of the alarm. Alarms can also be disabled so they don’t have to be deleted and recreated.

More improvements include:

  • Setting multiple alarms for CPU, memory, and disk space. For example, set a friendly reminder to Slack on a low threshold and when it gets close to maximum trigger PagerDuty or any other service.
  • Trigger each alarm from the UI with the ability to see how the alarms look and how it would react when triggered. This is very useful when trying webhooks.
  • The same goes for recipients - edit each recipient and trigger a notification for that recipient.
  • Give a recipient a custom name to make it easier to remember which one is which.
  • For VictorOps integration, it is now possible to use custom routing keys, which fulfills a popular user request.

If you have any questions or concerns regarding the improved alarms, send an email to Support@cloudkarafka.com.

CloudKarafka - Industry Leading Apache Kafka as a Service