Why would you use a scheduled job in bsa to monitor itself ? if the appserver fails, then how will you know if the job didn’t run ? you are talking about the bsa appserver right?
No ...I actually meant web server instances so it is not BL application server. i just want to know the impact of such frequent jobs in BL. It is just executing few commands to identify Tomcat Application on 5 target servers per instance. So If I schedule it for every 15 mins, will it create any impact related to performance?
Why do you want to use bsa to be a monitoring tool ?
How many targets are you going to ‘monitor’ w/ this job ? not sure what ‘5 target servers per instance’ means - how many instances do you have ? 10, 100, 1000 ?
It uses resources on the appserver that will not be available for other jobs.
The short answer is: no more than any other job would.
BSA is meant to execute jobs, and all jobs run using a schedule (in the background), even ad-hoc runs you do manually, so the scheduling of a job every 15 minutes on itself will not affect the performance of the app server more than any other job run manually if that's what you're asking.
Then it's a matter of checking what capacity you have in terms of concurrent NSH threads and work item threads and such. If you schedule a job every 15 minutes and it takes a while to run, it will lock a thread or more (depending on the type of job) for that much time and prevent other jobs from using it. If you end up with too many scheduled jobs that run every 15 minutes, you risk having a capacity issue eventually if you're also very active with manual job runs of all sorts in the console.
Bill, We do not have any scheduling tool in our environment so using Bladelogic. We are executing a script to check status of application process and as Yanick mentioned this will create an impact on the number of WITs and these could get blocked if job takes long durations to complete.
I believe database threads will also get consumed and if another job needs more threads then it will fail.
when you run a job, it will use a wit for each target. each wit will use up to 2 db connections from the jobexecutionconnection pool. the wit to db connection ratio should already be set properly so you would run out of wit before db connections.
i'm not asking about a scheduling tool, i'm asking about a monitoring tool. do you not have any monitoring tools in your env ?
how many targets will this run against ? you said 5 hosts per instance but didn't say how many instances.
this will consume as many wit in your env as there are targets, for however long it takes the check to run.
Instance count is one per server and we do have monitoring tool BPPM as well in the environment but we are monitoring limited parameters and we are looking forward to explore options through BPPM as well.
However, I would like to know how should I decide on the frequency of scheduled job to ensure not all WITs are consumed.
Functionality to monitor WorkItemThreads is available with the the BSA KM in BMC Truesight (BPPM)
How were you thinking of doing it via BSA?
So you have 5 servers in total ?
yes...and job takes less than a minute to complete