I would guess the easiest way (but not the most accurate way) would be to monitor the logs instead
To make it more accurate you can moniter the arserverd procces, and if that is up, then only check the logs (The arserver daemon can die without writing anything to the logs)
I would check the armonitor.log and email log files to see if something died after starting up (meaning what is the last state of a process)
Hope this would help seeing that you wont need any API's to do this and you would be able to extend this.
An easy way to monitor that your application server and mid-tier is up and functioning is to call a simple web service programmatically.
2 of 2 people found this helpful
Create a user with no permissions or licenses. Only create then in the user form and not CTM:People. So it doesnt matter if people find out the password and try and do something. Create a form with a lable on saying something like working. Again, not in any application or any special permissions needed. Then you can use a direct url to that form, passing the username, password, formname etc
Not 100% sure the above syntax url is correct but you get what I am saying.
Then use curl to do a get and you could grep for that label text in the html.
Column Technologies makes a product called Sentinel that monitors the AR processes, log files, events, applications, email timing, delays, ticket creation/modifications and other useful information.
It covers the ARServer, Mid-Tier layers and DB layers and has a web interface along with alerting etc.
Let me know if you want more information.
I guess Danny is getting sentimental here....
A customer where I did some work had a pretty good solution with software from HP actually logging in to the mid-tier and opening the incident management console, generating reports on if it succeeds and how long it took. Which was handy at a few occasions because you'd have instant benchmarks! I guess it was a pretty expensive piece of software though.
I might recomment to log in to the mid-tier and do some stuff to check Remedy's state as it will use lots of components, as you already noted.
You might also find this old thread on the ARSList interesting:
Like a grandfather sitting in his rocking chair talking about the good old days
Well if you want to go out and buy something that logs in and do stuff, why not just get the BMC TM ART ?
It does what Michiel explained
BMC TM ART is not exactly a cheap product - very good though.
Used in conjunction with other products you can build a full system monitoring solution, but it becomes an expensive exercise for the smaller companies.
* My Sales Mode Here *
I mentioned Sentinel from Column, as it is an affordable solution that will perform actions that currently requires more than one tool to achieve.
It is aimed to be affordable for both the smaller and larger companies and offeres a full solution without the need for more hardware overheads.
* End of Sales Pitch *
Wow, that's great to see all responses in such a short window. You guys are really amazing. I thought about checking the logs and wev service and to some extent I think the logs are useful sources to make the monitoring process more robust, but this is something I can build on later. Anyway, what dannykellett mentioned above seems to be interesting and I never thought about it. I'll give it a shot and see. Thanks again for all your inputs.