fix bookmarks and add disk pressure fixes

Author: rsgel

articles/chaos-studio/chaos-studio-fault-library.md

@@ -128,7 +128,7 @@ Agent-based faults are injected into **Azure Virtual Machines** or **Virtual Mac
 |-----------------|------------|
 | Load | [Start load test (Azure Load Testing)](#start-load-test-azure-load-testing) |
 | Load | [Stop load test (Azure Load Testing)](#stop-load-test-azure-load-testing) |
-| Time delay | [Delay](#time-delay) |
+| Time delay | [Delay](#delay) |
 
 ## Details: Agent-based faults
 
@@ -606,14 +606,16 @@ Currently, the Windows agent doesn't reduce memory pressure when other applicati
 | Prerequisites | **Linux**: The **stress-ng** utility needs to be installed. Installation happens automatically as part of agent installation, using the default package manager, on several operating systems including Debian-based (like Ubuntu), Red Hat Enterprise Linux, and OpenSUSE. For other distributions, including Azure Linux, you must install **stress-ng** manually. For more information, see the [upstream project repository](https://github.com/ColinIanKing/stress-ng). |
 | Urn | urn:csci:microsoft:agent:linuxDiskIOPressure/1.1 |
 | Parameters (key, value) |  |
-| workerCount | Number of worker processes to run. Setting `workerCount` to 0 generated as many worker processes as there are number of processors. |
-| fileSizePerWorker | Size of the temporary file that a worker performs I/O operations against. Integer plus a unit in bytes (b), kilobytes (k), megabytes (m), or gigabytes (g) (for example, 4 m for 4 megabytes and 256 g for 256 gigabytes). |
-| blockSize | Block size to be used for disk I/O operations, capped at 4 megabytes. Integer plus a unit in bytes, kilobytes, or megabytes (for example, 512 k for 512 kilobytes). |
-| targetTempDirectory | (Optional) The directory to use for applying disk pressure. For example, "/tmp/". If the parameter is not included, pressure is added to the primary disk. |
+| workerCount | Number of worker processes to run. Setting `workerCount` to 0 generates as many worker processes as there are number of processors. |
+| fileSizePerWorker | Size of the temporary file that a worker performs I/O operations against. Integer plus a unit in bytes (b), kilobytes (k), megabytes (m), or gigabytes (g) (for example, `4m` for 4 megabytes and `256g` for 256 gigabytes). |
+| blockSize | Block size to be used for disk I/O operations, greater than 1 byte and less than 4 megabytes (maximum value is `4095k`). Integer plus a unit in bytes, kilobytes, or megabytes (for example, `512k` for 512 kilobytes). |
+| targetTempDirectory | (Optional) The directory to use for applying disk pressure. For example, `/tmp/`. If the parameter is not included, pressure is added to the primary disk. |
 | virtualMachineScaleSetInstances | An array of instance IDs when you apply this fault to a virtual machine scale set. Required for virtual machine scale sets in uniform orchestration mode. [Learn more about instance IDs](../virtual-machine-scale-sets/virtual-machine-scale-sets-instance-ids.md#scale-set-instance-id-for-uniform-orchestration-mode). |
 
 #### Sample JSON
 
+These sample values produced ~100% disk pressure when tested on a `Standard_D2s_v3` virtual machine with Premium SSD LRS. A large fileSizePerWorker and smaller blockSize help stress the disk fully.
+
 ```json
 {
   "name": "branchOne",
@@ -624,23 +626,19 @@ Currently, the Windows agent doesn't reduce memory pressure when other applicati
       "parameters": [
         {
           "key": "workerCount",
-          "value": "0"
+          "value": "4"
         },
         {
           "key": "fileSizePerWorker",
-          "value": "512m"
+          "value": "2g"
         },
         {
           "key": "blockSize",
-          "value": "256k"
+          "value": "64k"
         },
         {
           "key": "targetTempDirectory",
           "value": "/tmp/"
-        },
-        {
-          "key": "virtualMachineScaleSetInstances",
-          "value": "[0,1,2]"
         }
       ],
       "duration": "PT10M",
@@ -650,6 +648,7 @@ Currently, the Windows agent doesn't reduce memory pressure when other applicati
 }
 ```
 
+
 ### Stop Service
 
 | Property | Value |

articles/chaos-studio/chaos-studio-tutorial-dynamic-target-cli.md

@@ -57,7 +57,7 @@ You've now successfully added your virtual machine scale set to Chaos Studio.
 
 Now you can create your experiment. A chaos experiment defines the actions you want to take against target resources. The actions are organized and run in sequential steps. The chaos experiment also defines the actions you want to take against branches, which run in parallel.
 
-1. Formulate your experiment JSON starting with the following [Virtual Machine Scale Sets Shutdown 2.0](chaos-studio-fault-library.md#version-20) JSON sample. Modify the JSON to correspond to the experiment you want to run by using the [Create Experiment API](/rest/api/chaosstudio/experiments/create-or-update) and the [fault library](chaos-studio-fault-library.md). At this time, dynamic targeting is only available with the Virtual Machine Scale Sets Shutdown 2.0 fault and can only filter on availability zones.
+1. Formulate your experiment JSON starting with the following [Virtual Machine Scale Sets Shutdown 2.0](chaos-studio-fault-library.md#vmss-shutdown-version-20) JSON sample. Modify the JSON to correspond to the experiment you want to run by using the [Create Experiment API](/rest/api/chaosstudio/experiments/create-or-update) and the [fault library](chaos-studio-fault-library.md). At this time, dynamic targeting is only available with the Virtual Machine Scale Sets Shutdown 2.0 fault and can only filter on availability zones.
 
     - Use the `filter` element to configure the list of Azure availability zones to filter targets by. If you don't provide a `filter`, the fault shuts down all instances in the virtual machine scale set.
     - The experiment targets all Virtual Machine Scale Sets instances in the specified zones.

articles/chaos-studio/troubleshooting.md

@@ -53,7 +53,7 @@ Some problems are caused by missing prerequisites.
 
 ### Agent-based faults fail on a virtual machine
 Agent-based faults might fail for various reasons related to missing prerequisites:
-* On Linux VMs, the [CPU Pressure](chaos-studio-fault-library.md#cpu-pressure), [Physical Memory Pressure](chaos-studio-fault-library.md#physical-memory-pressure), [Disk I/O pressure](chaos-studio-fault-library.md#disk-io-pressure-linux), and [Arbitrary Stress-ng Stress](chaos-studio-fault-library.md#arbitrary-stress-ng-stress) faults all require that the [stress-ng utility](https://wiki.ubuntu.com/Kernel/Reference/stress-ng) is installed on your VM. For more information on how to install stress-ng, see the fault prerequisite sections.
+* On Linux VMs, the [CPU Pressure](chaos-studio-fault-library.md#cpu-pressure), [Physical Memory Pressure](chaos-studio-fault-library.md#physical-memory-pressure), [Disk I/O pressure](chaos-studio-fault-library.md#linux-disk-io-pressure), and [Arbitrary Stress-ng Stress](chaos-studio-fault-library.md#arbitrary-stress-ng-stressor) faults all require that the [stress-ng utility](https://wiki.ubuntu.com/Kernel/Reference/stress-ng) is installed on your VM. For more information on how to install stress-ng, see the fault prerequisite sections.
 * On either Linux or Windows VMs, the user-assigned managed identity provided during agent-based target enablement must also be added to the VM.
 * On either Linux or Windows VMs, the system-assigned managed identity for the experiment must be granted the Reader role on the VM. (Seemingly elevated roles like Virtual Machine Contributor don't include the \*/Read operation that's necessary for the Chaos Studio agent to read the microsoft-agent target proxy resource on the VM.)