status of dSource

  • 0
  • 1
  • Question
  • Updated 7 months ago
  • Answered
Recently we had an issue where our delphix engine hung and as a result, our dSources were a little behind in the logs with our production database. Is there a way to proactively monitor the delphix engine and dsource so we can get notified sooner rather than when we find out manually ?
Photo of Srividya Srinivasan

Srividya Srinivasan

  • 304 Points 250 badge 2x thumb

Posted 7 months ago

  • 0
  • 1
Photo of Sathees Kodi

Sathees Kodi, Employee

  • 230 Points 100 badge 2x thumb
Yes. You can setup a alert profile and start receiving an email about the warning/critical faults. So that you will be notified as and when log file is missing.
Photo of Srividya Srinivasan

Srividya Srinivasan

  • 304 Points 250 badge 2x thumb
Thank you. Is the alert profile setup done from the GUI ? I haven't done it before. so just checking
Photo of Sathees Kodi

Sathees Kodi, Employee

  • 230 Points 100 badge 2x thumb
Not GUI. You have to use CLI. Here is the link for setting up alert profile.

https://docs.delphix.com/display/DOCS50/CLI+Cookbook%3A+Creating+Alert+Profiles
Photo of Srividya Srinivasan

Srividya Srinivasan

  • 304 Points 250 badge 2x thumb
Thank you very much. Does this also monitor the status of the delphix engine or rather report if there is a connection failure to the engine ?
Photo of Mouhssine SAIDI

Mouhssine SAIDI

  • 4,732 Points 4k badge 2x thumb
Hi Srividya,

You have several methods for monitoring the engine.

Using standard SNMP

SMTP for sending emails

Syslog to send Delphix logs to a syslog server

Custom scripts using our CLI or API, dxtoolkit

Mission control (Delphix reporting) - Managing multiple Delphix Engines

Regards,

Mouhssine
Photo of Ranzo Taylor

Ranzo Taylor, Employee

  • 1,572 Points 1k badge 2x thumb
All good points above, but only works if the Delphix Management component is working.  What if it's down?  It cannot send alerts.

 In addition to configuring alerts and notifications, you should use a simple HTTP/HTTPS probe from your monitoring solution of choice.  They all have such a probe, and your company is almost certainly already using one to monitor other important web based software, you just need to find the right people. 

The probe typically just hits a web page every minute and looks for a 200 response from a web server, and as long as they get one everything is judged to be OK.  It's very simple, and about 99% effective.

The reason only 99% effective is false positives and negatives.  False Positive:  the Management component might just be slow, not down.  (Personally...I'd like to get alerted for that).  False Negative:  hypothetically the Management component might be up but the engine is still "unhealthy".  (But you should get an Alert for that!)

Just remember, you need BOTH the HTTP/HTTPS probe AND the Delphix alert system to successfully monitor your engine health.