Forum Discussion

dcharling's avatar
dcharling
Data Storage
10-21-2021

How do you administer your Sisense installation?

I've struggled to find answers/solutions to some of these items:

  1. I need notice of bad performance and system issues before a user sends me an email.
  2. what errors are we getting? For example:
    1. Did any scheduled reports fail today?
    2. Did a user access a Dashboard that failed to render?
  3. how long do various builds take?
  4. what kind of resources are we using?
  5. how do we clean up old users that have not used the system in x days?
  6. how many dashboards do people have?
  7. how do we clean up old dashboards?
  8. what data can users access?
  9. what tables and columns do dashboards use?
  10. What are the usage patterns for our dashboards? 
    1. which ones aren't being used anymore? 
Technology

2 Replies

Replies have been turned off for this discussion
  • irismaessen's avatar
    irismaessen
    Data Pipeline
    10-25-2021

    Sisense Monitor and Sisense Usage Analytics provide quite a few useful overviews and roads to investigations, but one downside I've found is that the possibilities for alerting on these metrics is limited. You could use Pulse to send alerts based on average dashboard load time, but if you want to be proactive with this, that means rebuilding the analytics cube every x minutes which is less than ideal.

    I like the Usage Analytics for the information about dashboard usage and overall dashboard performance, but I'm missing information there about overall performance of individual widgets. That information is available in the Monitor on an individual-query basis, but there the behaviour of widgets over time (as opposed to all queries on a server) is still difficult to analyse.

     

    Perhaps on Linux the (very nice-looking) Prometheus+Grafana dashboards provide better handles for performance-related alerting. Does anyone have experience with that already?