Monitor Losing Connection

Andrew_Cant

Since upgrading to the 2026 version of monitor (currently on 26.1), we've noticed that Monitor will lose connection to our servers a few times a day.

Each disconnect will last approx 5 mins, before reconnecting without issue.

There is no pattern behind this, it will happen seemingly at random times of day and night.

This appears to be a monitor issue as we have manually checked the network, and each environment at the time when monitor has lost connection, and all appears to be fine.

This will happen on Portal, Enterprise Server, Hosting Server, Image Server, Hosting Server and Web Server, but happens more on Portal and Enterprise Server. 

Checking the logs, I can see no reason that this is happening. This is both the Monitor logs, and also the server logs.

Can anyone suggest what may be happening?

Many thanks

Andrew

Pam-Pritchett

I’m seeing very similar behavior, although our environment is still running ArcGIS Monitor Server 2025.1 with ArcGIS Monitor Agent 2025.1.1, not 2026.1.

Monitor has intermittently lost connectivity to multiple otherwise healthy components, including ArcGIS Server, the relational Data Store, and SQL Server. This generates a large number of critical connectivity alerts even though checks of the underlying servers, services, CPU, memory, and network have not shown an actual outage.

During one occurrence, the ArcGIS Monitor Server and Agent services both remained running and port 30443 was still listening, but Monitor stopped completing application requests. It later recovered on its own. During another occurrence, the Monitor Server service had to be restarted.

In several cases, the behaviour began while or shortly after opening Administration → Logs. The Logs page can take 2–3 minutes to load, and during that time Monitor has generated widespread connectivity alerts and, on some occasions, become unresponsive. I have not confirmed that accessing the Logs page is the root cause, but it appears to be a repeatable trigger or contributing factor in our environment.

We have increased the cluster size from 2 to 4 and returned logging to Warning, but the behaviour has continued. I currently have Esri Support case 04221248 open and will share any useful findings from the investigation.