1

We have used Fogbugz for years (since v4, currently running v6). A few times a year we will find that the Fogbugz Maintenance service has stopped.

It is "no big deal" in that we just start it up, and all is well.

It is a big deal because we usually (for example, yesterday) will lose a day of work because some user does not live in his FB list, and did not get an email of new (urgent) cases being assigned to him.

OK, so that is the situation.

The question is: How can we monitor that the FB maintain service is running?

On the web software we make, we have monitoring URL's, and we point a web monitor service at those urls. When the urls time out, or return errors, the monitor service sends an SMS message to the phones of the two ops guys, and they jump on it. (Our web app has services like the FB maint service, and the monitor urls also are able to tell if those services are alive.)

Can we do this in FB?

(Let us know if this is in FB6, 7 or both... we need to know these things...)

Thanks!

Fog Creek Case FC1838018

flag

4 Answers

3

Thanks. I know I can use MS MOM, Hyperic, Polymon, etc to monitor services. But we do all our critical monitoring with an off site, by subscription service. We don't have enough ops internally to do the whole IT monitoring story in house. – Sam Jones 3 hours ago

Then I believe you would have to create a plugin and in that plug in you can execute the select like David said. Then you can see maybe in the extra menu a option like this 'See Maintenance Service Status' and show the result of the query.

link|flag
That will work! – Michael Pryor Sep 14 2009 at 15:45
2

If you don't mind giving your service database access, you could query

SELECT sValue FROM Setting WHERE sKey = 'dtLastHeartbeat'

If that date/time is greater than a few minutes ago, the maintenance service is not running correctly.

link|flag
That doesn't help me monitor fb via a url... Gotsta be a url. I am not doing the whole "service monitoring a service" thing.... – Sam Jones Sep 11 2009 at 16:12
0

A cron job running a standard query against the MailQueue table should do the trick, but there are a few wrinkles here. Every time someone performs an action that should notify another user, a row gets inserted into the MailQueue table.

A number that only goes up, rather than fluctuating (and hopefully reaching zero) as mail piles up and gets sent out, would indicate that maintenance is not running correctly (or that something had happened with the SMTP server).

FogBugz should warn admins if the maintenance service has not run in 45 minutes. It does this by putting a row in the Notification table, which will be removed once the maintenance service completes successfully. You can set up monitoring scripts against the FogBugz web site, and against the Notifications table.

This does not apply to FogBugz On Demand. For that product, we have our own set of monitoring scripts.

link|flag
Thanks, but I am looking to leverage our existing monitoring setup, which is url based. Adding chron jobs means I need to then start monitoring the chron jobs... (ack) Please consider this as a feature request for FB. With FB now, my monitor can tell that FB is up (cause the home page comes up). But it can't tell if FB is fully working (cause we can't monitor the service via a url). Thanks! – Sam Jones Sep 10 2009 at 18:58
0

We use a product call PolyMon that can monitor Windows Services. I totally recomended it.

link|flag
Thanks. I know I can use MS MOM, Hyperic, Polymon, etc to monitor services. But we do all our critical monitoring with an off site, by subscription service. We don't have enough ops internally to do the whole IT monitoring story in house. – Sam Jones Sep 11 2009 at 16:13

Your Answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.