How can we help you today? How can we help you today?

Are thresholds configurable by user?

How are the thresholds set for things like CPU and Free disk drive space.

I have different levels of servers and I don't care if dev uses 95% of disk drive but production should not go above 80%.

Same with CPU so servers for me are like different classes just like you have different levels of monitoring..

Will you be able to set a threshold at the server level in the final product?

-jfc-
johnc650
0

Comments

6 comments

  • Dan J Archer
    Hi John,

    A very good question, and one we're still debating. We'd like to have more feedback from our beta users to understand what people need in this area, and why.

    One available solution in the beta is that if you really aren't worried by the disk space consumed by your dev machines, then after the disk space incidents come up you can hit "Ignore on <MACHINENAME>" to prevent them from coming up on that machine in the future.

    Thanks,
    Dan
    Dan J Archer
    0
  • Giggles220
    I agree with John. I have different requirements on different boxes, I could very easily categorize my boxes in the 4 levels you have and set a different theshold for those categories though.
    Unless I missed it I don't see anyway to even set the threshold globally for all servers.
    Giggles220
    0
  • johnc650
    Different levels of thresholds for example on Prod boxes.. CPU

    LOW ALERT CPU >= 75%
    MED ALERT CPU >= 80
    HIGH ALERT CPU >= 90

    They show up on the web page as YELLOW, PINK AND RED.

    -jfc-
    johnc650
    0
  • johnc650
    Just got this alert: The log file for database MDS is 136.49MB and has 134.28MB of free space (98.38%). It is recommended to shrink the file to avoid excess wasted space..

    But who cares we are talking 135MB here... but I can't set a threshold for this DB or turn off for this DB.. I can only say ignore at the SERVER level or EVERYWHERE... If I ignore on this server the OTHER 35 DBs on it will now NO LONGER monitor for wasted space...CORRECT??


    -jfc-
    johnc650
    0
  • broughtie
    I currently use SQL Sentry to monitor our SQL servers, and although the product has a number of nice features, the section that monitors SQL Agent jobs is the part I use the most and if RedGate SQL Response if powerful enough, I'd probably dump SQL Sentry for SQL Response.

    If if only managing a single server, I still see the ability to configure incidents as one of the most important features. In the current beta of SQL Response, its a bit of a mystery what "incidents" are even being monitored. It would be nice to be able to pull up a list of all possible incidents and be able to configure each one (turn on /off, set default monitoring thresholds, etc).

    Specific to the topic of this thread, I thing every incident type need to be configurable for each object covered by that incident type. For example, lets look at the "Long Running Query" incident. Lets say I have three jobs setup in SQL Agent that execute some code. Job 1 runs continuously and never ends, Job 2 runs for about 3 seconds every 10 minutes, and Job 3 runs for about 15 minutes once a day. Now for each of these jobs, I'm going to want to configure individually how "Long Running Query" is defined. Job 1 shouldn't even be included in this incident type, Job 2 should be considered long running if it runs for more than 30 seconds, and Job 3 shoudl be considered long running if it runs for more than 20 minutes. Now no one is going to want to have to configure each incident for every object as part of the setup since that would be daunting in many environments, and the is where the concept of default values for an incident type come into play. Perhaps I can define "long running query" to be anything that take more than 10% of the average execution time. Or maybe anything that runs for more than 30 seconds. These defaults might cover 90% of my objects, and I could then go into individual objects and make specific adjustments for those objects that differ from the default.
    broughtie
    0
  • BEACHDBA
    I would absolutely be interested in configurable thresholds and allowing for different thresholds for different server types. Since monitoring levels is a different beast than thresholds, I think thresholds should be kept separate. If I could generate threshold templates that could be named (e.g., dev, test, prod) and associate each server with such a template, that would be awesome. If I could override the templates, per server, as needed, that might be too cool. :shock:
    BEACHDBA
    0

Add comment

Please sign in to leave a comment.