We had the webinar giving an overview of the new Performance Monitoring Toolset last week. Learn about the end-to-end performance management right from the connector, application, individual lines of code, databases, incoming/outgoing services, system metrics, memory management to JVM metrics.
The recording is available at https://my.adobeconnect.com/p70rndpsrny.
Join now for part 2 of the webinar series on July 17th to understand the various scenarios that you might encounter while working with your CF application. This include issues around JVM, database issues, application specific issues and issues with system architecture. Register now at https://meetus.adobeevents.com/coldfusion/
Staff
55 posts
When installing PMT on a separate server, does the OS of the PMT server have to match that of the CF server (PMT:Windows/CF:Windows), or can it be different (PMT:Linux/CF:Windows)?
Yep, Paul. It’s a common problem. It’s not that you CANNOT install the PMT on the same machine as CF, but you will nearly always find you should not. It is a beefy app (the PMT and the datastore, ElasticSearch), so for most it would generally be best to be put on its own machine, so as not to so directly impact the one being monitored.
This was mentioned in the video (though perhaps you have not watched it, or you may have missed it), but they also mentioned it in many of the initial demos of it. Most recently, the docs have been updated to add a section to help plan its deployment (https://helpx.adobe.com/coldfusion/performance-monitoring-toolset/deployment-guide.html). As it appears now as the “last” section of the PMT docs, they would do well to point the first “installing” section to this, for more.
Sorry, no, I have not yet watched the video yet as we were only just made aware of it. But, we quickly figured out we couldn’t run it on our CF servers due to the resources required. We don’t want to run it on any of our other servers for that matter. We’re running everything in the cloud, and the prospect of paying for another VM just to run the PMT is a bit much.
We tried running it from our office and connecting to the CF servers through a VPN but the PMT was having none of that. We were not able to get PMT to ‘see’ the servers over our VPN.
So….we’ve left PMT installed on one of our VMs but not running, figuring we can start it up should the need arise.
We’ve begun looking at third parties for every day monitoring such as Fusion Reactor and SeeFusion. We’re used to having the quick peek at usage provided in the previous CFAdmin monitor from CF11.
Yep, you are not alone in any of those observations and conclusions.
We were looking forward to CF2018’s Performance Monitoring Toolset when we recently updated our servers. However, the base install of the monitor and datastore was consuming over 2.5GB RAM and bogging down any machine (Windows) we tried to install it on. Anyone else run into this?
Hi Paul,
We understand your concern with the default memory consumption. However, PMT has an option to configure the memory consumption according the deployment needs. You can easily configure PMT to run with 512 MB but you might see little lag in rendering the PMT dashboard. If PMT datastore has more information then it can retain in-memory.
We also have a PMT deployment guide that Charlie shared in the previous comment. You can follow the instructions given the guide to reduce the memory footprint as per your need. PMT deployment guide also talks about archiving strategies.
Here’s the link to the deployment guide:https://helpx.adobe.com/coldfusion/performance-monitoring-toolset/deployment-guide.html
Please reach out to me @ nimsharm@adobe.com if you face any issue.
-Nimit
You must be logged in to post a comment.
23 Nov
10AM Pacific
Online
