updates2

Author: LiSeda

articles/cyclecloud/release-notes/8-2-1.md

@@ -8,19 +8,17 @@ ms.author: adjohnso
 
 # CycleCloud version 8.2.1
 
-## New Features:
+## New features:
 
-* Slurm hostnames now match their node names, via Azure DNS
-* [VMSS Force Delete](~/articles/cyclecloud/how-to/terminate-cluster.md#force-delete-virtual-machines) is supported (if enabled)
-
-## Resolved Issues:
-
-* CycleCloud would sometimes get into a mode where connections failed with `__getattribute__ not found on type HttpsURLConnectionImpl` until it was restarted
-* When editing accounts, the credential's default region was erroneously reset to WestUS.
-* In some cases, CycleCloud requested Azure metrics for an interval that was too large and got errors
-* Spot VM evictions that occurred while VMs were booting were not handled properly
-* Certain VM errors were incorrectly classified as a VM error rather than a no-spot-capacity indicator
-* Catching up on metrics could cause an out-of-memory error if CycleCloud had been off for an extended time
-* Various accessibility improvements
+* Slurm hostnames now match their node names via Azure DNS.
+* [Force-delete virtual machines](~/articles/cyclecloud/how-to/terminate-cluster.md#force-delete-virtual-machines) is supported (if enabled).
 
+## Resolved issues:
 
+* CycleCloud sometimes got into a mode where connections failed with `__getattribute__ not found on type HttpsURLConnectionImpl` until it was restarted.
+* When editing accounts, the credential's default region erroneously reset to WestUS.
+* In some cases, CycleCloud requested Azure metrics for an interval that was too large and got errors.
+* Spot virtual machine evictions occurring while VMs were booting weren't handled properly.
+* Certain VM errors were classified incorrectly as a VM error rather than as a no-spot-capacity indicator.
+* Catching up on metrics could cause an out-of-memory error if CycleCloud had been off for an extended time.
+* Various accessibility improvements