PDQ Inventory 16 Beta2/3 so far
Just some ramblings on the current beta.
Pressing F5 seems to wipe out any applied filters now.
The agent seems to sporadically go into an Unavailable state and become unresponsive after a reboot. We have been reinstalling the agent to fix this (pre Beta 3) Following our update to Beta 3 we are updating to the new agent to see if this continues.
Installing the agent on multiple computers has a solid chance of making your server temporarily unresponsive. Inventory probably needs to share the max scan setting for agent deployments. Anything over ~200 at once here causes server disconnect.
Having the agent live update user log on and off is a godsend.
I'm wishing for some kind of merge of Agent status and scan status just from a screenspace point of view.
How is everyone else doing with the current beta?
Comments
I have a few questions you might be able to answer since the only documentation on the agent is how to install/uninstall it.
What exact does the Agent do in it's current form?
What benefits have you noticed so far?
How exactly does it connect to the inventory server?
Is it pull (from the server) only or does it push anything to the server (scan details?) on a regular basis to the server?
Thanks!
Just have not seen any details on exactly what the Agent includes.
Just noticed this new link at the top of the page. I'll check there.
https://beta.pdq.com/
Not a Beta related note, but i hope maybe helpful for others.
I switched from "client mode" to "central server" to test new agent.
Central Server was never necessary for me until now (single admin) but after switching to central Server both, PDQ Deploy and PDQ Inventory, running much much smoother.
The GUI response faster, nothings hang, like back in the old days.
All the documentation is here https://documentation.pdq.com/PDQInventory/16.0.3.0/ and https://documentation.pdq.com/PDQInventory/16.0.3.0/index.html?agent.htm
The agent (from what I have read / noticed) attempts to connect to your instance of PDQ inventory and report noticed events (user log on/off startup/shutdown) it reports these within seconds to inventory. Handy for keeping currently logged on users accurate.
If the agent can't communicate with your instance of PDQ inventory it trys to talk to PDQ.com and I assume relays this somehow. We may need to find a way to disable this to keep within security compliance.
The major benefits are a more accurate heartbeat and better resiliency for DNS mishaps, and user log in accuracy. This seems like just a starting point as the agent has a great deal more potential.
I wonder if we will have the option to limit these events? With 25,000 endpoints and students logging in and out of computers multiple times a day I can see this becoming a huge drain on the server.
What I'm really hoping for is a way for the agent to regularly (at an admin controlled timeframe) connect to the server and report it's scan details instead of having to have the server reach out to the endpoints for every scan. We do this with Desktop Central and it's inventory is much more up to date for large networks like ours.
Oh, and thanks for the links but all the Doc's show are how to uninstall and install the agent. Nothing about what it does or how it communicates.
Small update, Beta 3 agents seem to go Unavailable at the same rate as before, we are also noticing several computers stuck in some kind of scan que for an rather long time. I aborted all scans yesterday, walking in this morning i see some that have been on hold for 18 hours waiting on agent.
update to the update, took a look at our PDQ server. CPU was slammed at 80-90% (6 vCPUs at 2.8ghz) bumped it up to 12 and holy cow things started moving. blasting through 200 scans like it's nothing and agents started reporting in and dropping the Unavailable status.