Tool to monitor HM / RMA

JuergenF · Post by **JuergenF** » Mon Apr 14, 2008 11:26 am

@greyhat64

I fully agree to you.
What I wanted to say is

JuergenF wrote:You can use Servers Alive for example

... as a workaround until it is implemented in HostMonitor Package

KS-Soft · Post by **KS-Soft** » Mon Apr 14, 2008 11:32 am

Probably we can provide 2nd license for 2-3 test items at no cost for everyone who purchased Enteprise or Professional license.
If you are interested please send request to sales@ks-soft.net. Please provide your order number and/or registration name.

Regards
Alex

greyhat64 · Post by **greyhat64** » Mon Apr 14, 2008 12:58 pm

KS-Soft wrote:. . . 3rd option is superfluous.

The reason I suggest that RMA alarm independently is twofold,
(1) This suggestion is based on the assumption that watchdog is a function of RMA. A watchdog alarm is necessary.
(2) In a situation where HM is not responding you may also have a critical alarm that RMA is trying to post. I may need to respond to the RMA side critical alarm BEFORE I respond to the HM not responding. Even if it were presented as a generic "Critical alarm can't post to HM" alarm I have improved my ability to prioritize.
There is nothing superfluous about that.

greyhat64 · Post by **greyhat64** » Mon Apr 14, 2008 1:08 pm

Of course I guess it could be argued that if I'm that concerned, simply buy a second license and have the HM's monitor each other.

Which is what you are offering here:

KS-Soft wrote:Probably we can provide 2nd license for 2-3 test items at no cost for everyone who purchased Enteprise or Professional license.
If you are interested please send request to sales@ks-soft.net. Please provide your order number and/or registration name.

Regards
Alex

Alex, you guys are absolutely unparalleled in your responsiveness to customer needs! Awesome!
(How many more superlatives should I extol?)

KS-Soft · Post by **KS-Soft** » Mon Apr 14, 2008 1:32 pm

(1) This suggestion is based on the assumption that watchdog is a function of RMA. A watchdog alarm is necessary.
(2) In a situation where HM is not responding you may also have a critical alarm that RMA is trying to post. I may need to respond to the RMA side critical alarm BEFORE I respond to the HM not responding. Even if it were presented as a generic "Critical alarm can't post to HM" alarm I have improved my ability to prioritize.

It has sense hower
1) I think watchdog should be a separate utility, it may provide several methods of HostMonitor monitoring, it may check event log to distinguish normal shutdown from crash and so on
2) I think "HostMonitor does not respond" alarm has high priority because you don't know what problems was covered by this event (how many problems do you have and what kind).

Alex, you guys are absolutely unparalleled in your responsiveness to customer needs! Awesome!

You are welcome

Regards
Alex

greyhat64 · Post by **greyhat64** » Thu Jun 26, 2008 11:54 am

Alex,
I had more thoughts regarding redundancy & failover:
Ataude's suggestion: for rendundancy seems workable, but not as elegant as including this as an option/feature of the product. And I see an opportunity to address two wishlist items in one.

It goes without saying that the second 'standby' HM would need to replicate the primaries settings and be ready to pick up where the primary left off.
Using this premise, regarding Wabiloo's request to transfer files from an Active RMA system to the HM, couldn't you potentially use the same codebase for both purposes, as they are both essentially file transfers?

Then there is a question regarding logging, especially via SQL. You were certainly forward thinking by including the checkbox option "Execute alert when log inaccessible".
Can this be used to post transactions to an alternate, i.e. text, log and have them posted to the primary log when it becomes accessible?
The reason for this need may be obvious: Gaps in your data bring into question the validity of the entire dataset and in order for Log Analyzer to have real value this would need to be addressed.

Another web enhancement opportunity:
It would be incredibly informative to post logging info - enabled/disabled, log type, & timestamp of the last posted entry - for the default log in the header of the HM Web Interface.
And for each test, a checkbox column if logging is enabled for that specific test, allowing an admin to turn on/off default logging from the web interface would be helpful. (Grey it out if private logging enabled)
And since you already have Private log: info on the details page, why alter that to show if it's a default or private log, and include the aforementioned fields.

I saw the post for 7.42 Beta. Keep up the good work!

KS-Soft · Post by **KS-Soft** » Thu Jun 26, 2008 12:57 pm

Using this premise, regarding Wabiloo's request to transfer files from an Active RMA system to the HM, couldn't you potentially use the same codebase for both purposes, as they are both essentially file transfers?

I think clustering version of HostMonitor should use more complicated algorithm. There is no much sense to transfer entire file when operator just disables or removes single test. Probably it will be close to HostMonitor <-> RCC protocol.

Then there is a question regarding logging, especially via SQL. You were certainly forward thinking by including the checkbox option "Execute alert when log inaccessible".
Can this be used to post transactions to an alternate, i.e. text, log and have them posted to the primary log when it becomes accessible?

That's why HostMonitor offers Primary and Backup logging options. When primary ODBC log is not accessible, HostMonitor uses backup log. There is no built-in option to merge logs because there is no standard log format (everyone using its own log tables) however you may easily create such procedure by yourself (just copy records from one database into other).
If we implement "standard" database format for logging functions in version 8 then we can modify Log Analyzer to work with both logs at the same time.

Another web enhancement opportunity:
It would be incredibly informative to post logging info - enabled/disabled, log type, & timestamp of the last posted entry - for the default log in the header of the HM Web Interface.

May be we provide option to display all settings (options), not just information about logs. Or may be we implement "templates for Web Service, in such case you will be able to create your own interface...

And for each test, a checkbox column if logging is enabled for that specific test, allowing an admin to turn on/off default logging from the web interface would be helpful. (Grey it out if private logging enabled)
And since you already have Private log: info on the details page, why alter that to show if it's a default or private log, and include the aforementioned fields.

H'm, not sure

Are you often changing logging options for various test items?

I saw the post for 7.42 Beta. Keep up the good work!

Thank you

Regards
Alex

greyhat64 · Post by **greyhat64** » Fri Jun 27, 2008 4:01 am

Alex,
Thanks for taking the time to break it down and answer as much as possible.

There is no much sense to transfer entire file when operator just disables or removes single test. Probably it will be close to HostMonitor <-> RCC protocol.

I see your point, but I see the logic to parse what gets transferred seperately from the code used to execute the 'push' of that data. My point being that, once you have a mechanism for transfer you could extend that to other functions, as in Wabiloo's request.

There is no built-in option to merge logs because there is no standard log format

An honest, but troubling response. My companies software caches log entries, posting them to the primary database once the connection is reestablished. It simply makes sense to consolidate the data if possible. If I were to rank my requests by personal priority this would be, at the very least, in the top three.

. . .may be we implement "templates for Web Service,

I like it! And while your at it, take the next step and include support other web servers.

Are you often changing logging options for various test items?

Of course not, I was just thinking on the fly - skip it. But could you include all the logging details on the test detail page, not just the Private log info?

KS-Soft · Post by **KS-Soft** » Fri Jun 27, 2008 5:45 am

My point being that, once you have a mechanism for transfer you could extend that to other functions, as in Wabiloo's request.

When I said "I do not think we will do that" I did not mean "its hard to implement". I meant "we should not do that". Purpose of RMA - reduce network traffic, increase security, avoid any file and data transfer except test results and so on. That's why it should not transfer any files to HostMonitor.

Of course not, I was just thinking on the fly - skip it. But could you include all the logging details on the test detail page, not just the Private log info?

Sure. I have added such task

Regards
Alex

Robert_in_MTL · Post by **Robert_in_MTL** » Tue Apr 21, 2009 1:07 pm

Hi Alex,
is this still in the TODO list?

thank you.

KS-Soft · Post by **KS-Soft** » Tue Apr 21, 2009 2:10 pm

If you own Enterprise license, we can provide 2nd HostMonitor license for several test items at no cost. So "watchdog" utility has low priority.

Regards
Alex

greyhat64 · Post by **greyhat64** » Tue Apr 21, 2009 6:00 pm

A second 'lite' license can be useful, but it is not the same as a watchdog utility. Enterprise clients deserve a little bit more than that.

In fact, an expanded 'RMA Manager' with host monitoring/alarming would probably be the ideal solution.

The Ideal Watchdog Utility:
An expanded 'RMA Manager' with host monitoring/alarming would be the ideal solution.
Ideally a watchdog utility is a bit more focused and 'intelligent' than an unconfigured 'lite' license. A true enterprise solution would be 'self monitoring' - it wouldn't be up to me to configure a suitable test set to determine if the primary server is up/down.

Similar to the way that RMA's communicate with the server via RMA Manager, a 'heartbeat' would be maintained, and even if there was a fundamental change on the server (say, IP address changes) the heartbeat would be maintained.
You might argue that additional 'lifesign' tests could be added or activated to confirm a host server's viability and to better diagnose the source of the problem. This could be provided through a checklist of common tests - ping, http, etc.
Similarly, alarm actions would be pre-configured during AHM installation when the administrator is registered on the host server. An administrator's profile, including primary notification method (email, IM, SMS, etc.) would be used, and changes to that profile would be communicated to the watchdog utility to ensure proper notification.

KS-Soft · Post by **KS-Soft** » Wed Apr 22, 2009 5:10 am

Enterprise clients deserve a little bit more than that

You want more functionality that can be provided by HostMonitor? Then such software should be called Extended HostMonitor

Watchdog utility cannot provide more than 60 test methods and 30 actions like HostMonitor does.
Also I don't think such utility should be similar to RMA Manager and I don't think we need another utility to monitor RMA. We can add some options to HostMonitor that works with RMA 24/7. HostMonitor knows if there is any RMA related problem, why we should invent and you should setup additional software?

Regards
Alex

Robert_in_MTL · Post by **Robert_in_MTL** » Wed Apr 22, 2009 6:30 am

To make things easier: could we just select 1 or 2 remote monitoring agents so that they would check on HostMonitor on a regular schedule (every 30 seconds or every minute, whatever) and if they cannot reach HM, they would send admins an alert email.

I woudn't mind configuring a RMA for an SMTP server and a mail Template (but the best would be to push those settings from RMA Manager)

You would need to enable a new version of RMA to monitor HM on it's own and enable it to send a pre defined mail on it's own. That's it!

We would then just have to go to options and select which agent(s) is(are) designated to be so-called watchdog(s), HMwould then watch them back.

In the end everyone would benefit from this, both KS-Soft and the users because for you this would be a simpler way to do it right?

Cheers and keep up the superb work !

(I appreciate HM every time I use it!)

KS-Soft · Post by **KS-Soft** » Wed Apr 22, 2009 7:05 am

H'm... I don't think its a good idea either.
What exactly is wrong with 2nd instance of HostMonitor that can be installed at no cost? Why you don't like it?
It provides several test methods that can be used to monitor main HostMonitor (TCP test, File/Folder Availability, SNMP Trap) and 30 actions!
Why additional tool with single action method is better??

Cheers and keep up the superb work

Thank you

Regards
Alex