Agent disconnected on Monitor 2024.1.0

[Oiligriv]

Hello
after a 99% disk saturation the postgres crashed.
after restarting the postgreSQL and doing some maintenance to the DB via arcgis monitor, everything started working again.
after 5 days I realized that all the agents were disconnected.
I did the following steps.
restarting the agent to see if it would reconnect by itself, instead it doesn't work.
launching the registration command specifying the id --agent-id but nothing it is not possible to reconnect the agent.

it seems similar to my old post, but in that case the agent id command worked

https://community.esri.com/t5/arcgis-monitor-questions/no-data-monitoring/m-p/1575252#M2404

Thanks 

Virgilio

[Oiligriv]

I was trying to view the arcgis monitor client logs, I don't know if it helps, but I noticed this statement

Max consecutive ping failures reached: disconnecting from server

Virgilio

[Oiligriv]

the registration was successful, I had recreated the admin user as viewer instead of admin.
but the mass disconnection anomaly remained

[DerekLaw]

Hi @Oiligriv,

Can you please contact Esri Tech Support and open a case so they can help investigate further? 

[JeffMertzXcel]

We also experienced a mass disconnection anomaly as you stated:-) We went to 2025.0 recently and are still in the process of cleaning things up and establishing new analysis pages, etc (we did not upgrade from 2024.1 but did a fresh install instead). Just noticed during a review that we "lost" 27 out of 38 agents. Restarting the agent process and/or monitor server didn't work. We ended up re-registering them with their respective agent id to get them communicating again. Agent logs also indicate many warning messages of "Ping task failed: Server may be unavailable or agent is disconnected. Reattempt in 60 seconds" log entries on several days eventually leading to the "Max consecutive ping failures reached: disconnecting from server" message. We have logged a ticket with Esri support to see if we can troubleshoot this more. The only thing what would prevent communicating with the server would be during Windows patching cycles, but these events do not directly correlate with that activity. Will post an update if/when we get to the resolution of this.