vSphere DRS Memory Overcommit Trigger Threshold Configuration

Answer

There is no fixed active‑memory percentage or absolute value that vSphere DRS uses as a universal high‑memory threshold. The threshold is calculated dynamically by the ESXi memory manager based on current active memory, swapped memory, host overhead, and workload characteristics, so it varies from host to host and can change over time.

Likely explanation (based on observed behavior)

In practice, administrators often see DRS start reclamation (ballooning, swapping, or vMotion) when host active memory rises into the 70‑80 % range of physical RAM. This is not a guaranteed limit; the actual point can be lower or higher depending on the Mem.MemMaxMax and Mem.SwapEnable settings and the specific VM memory demands.

How the advanced settings influence the threshold

1. Mem.MemMaxMax – defines the maximum memory a single VM can allocate relative to host physical memory (default = 100 %). Increasing this value lets individual VMs consume more host RAM before the host is considered over‑committed, effectively shifting the pressure point upward. Safe adjustment range is generally 100 %–200 % of host RAM; values above 200 % are rarely needed and risk host exhaustion.
2. Mem.SwapEnable – a boolean that enables or disables host‑level swapping of guest memory. When set to false, swapping is removed as a reclamation option, so DRS must rely on ballooning and vMotion alone. This makes the host appear under pressure earlier (the effective threshold moves downward). When true, swapping adds a reclamation path, which can delay the need for ballooning/vMotion. The only safe values are true or false; disabling swap should be done only after confirming sufficient physical RAM and adequate swap space on datastores.

Can the threshold be exposed or configured via API/CLI?

No. The exact memory‑pressure threshold used by DRS is internal to the ESXi kernel and is not exposed through any supported vSphere API, CLI command, or scripting interface. Attempting to read or alter it via undocumented options can violate VMware support terms and may destabilize the host.

Practical verification steps

1. Log into the ESXi host via SSH and run esxcli host memory get to view current active memory, swapped memory, and total RAM.
2. In the vSphere Client, open the Host Performance tab, select the Memory panel, and watch the Memory Pressure graph; the upward trend indicates where DRS begins to act.
3. Use esxcli system performance metric get -m cpu.memory.active to poll the live active‑memory percentage and compare it to the pressure point observed in the performance chart.

Missing diagnostic detail

To refine any version‑specific guidance, please confirm the ESXi version you are running (e.g., 6.7, 7.0, 8.0).