Scout Plugins |
Scout is a server monitoring & reporting application that's powered by plugins like this. Once you create your free Scout account, installing a plugin is a button-click away. |
|---|
Plugins » Ruby on Rails Request Monitoring
This Plugin URL has not been approved yet.
Monitors ALL requests to a Ruby on Rails web application, alerting when slow web requests occur.
Reports back the number of total requests, the number of slow requests, the average execution time, and the full URLs of slow requests.
Requires the elif library
sudo gem install elif
Options:
- Location of Ruby on Rails log file
- Maximum request length
9 Comments
| Installs: | 417 |
| Tested On: | linux macosx solaris |
| Source: |
View Source
All plugins listed in the directory are open source.
|
I’m using on Solaris
Rafi – updated to reflect Solaris, thx!
It would be nice to be able to specify urls that are not included in the slow requests. We have some URLs which we know take longer to execute and we’d like to not get constant notifications for those URLs.
We added support for the newer Rails configurations that work in miliseconds. This plugin still works in fractional seconds for backwards compatibiliy. Times from newer Rails logs will be converted.
@Brian – I agree. I believe the difficulty is that the line we are parsing to output the time doesn’t display the HTTP method used, so /users [POST] and /users [INDEX] would be interuppted as the same url. We are planning a fairly major overhaul to handle cases like this.
does this work with syslog-style logging (for http://nubyonrails.com/articles/a-hodel-3000-compliant-logger-for-the-rest-of-us)? doesn’t seem to…
I don’t quite understand the metrics this plugin monitors, and the lack of documentation does not help.
Lack of units and axis label in graphs is also an annoyance.
The graph’s left axis, does the highest value for any line get 100%, with rest of data shown appropriately?
If the above is right, all my requests are slow (defined to be 5secs+) as I am getting a constant 100% for slow requests. On hovering over this line, no popup with value is shown, unlike that of the other two lines.
Average request length: I am getting values mostly around 0.18. What does it mean? it obviously isn’t in seconds, as it’s reporting a lot of requests over 5 seconds.
Request count: is this average requests per hour?
More documentation per plugin explaining all the in’s and out’s is much needed.
I need to demo this to a client on Thursday, so can you please answer the above as a matter of urgency?
@Woody – You can install the syslog-version
@Xin
Sorry for any confusion.
> The graph’s left axis, does the highest value for any line get 100%,
> with rest of data shown appropriately?
By default, your data is scaled. This allows us to plot average requests, slow requests, and total requests on the same graph so you can look for correlations.
> Average request length: I am getting values mostly around 0.18. What does it mean?
> it obviously isn’t in seconds, as it’s reporting a lot of requests over 5 seconds.
The values are in seconds.
> Request count: is this average requests per hour?
This is the # of requests that were processing between client runs (so if the client runs every 10 minutes, it is the # of requests every 10 minutes).