#autoscaling — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #autoscaling, aggregated by home.social.
-
The hailstorm last week pushed several Swiss insurance claim portals to their limits. ⛈️
Portals were temporarily overloaded or unreachable.
With Managed Kubernetes (NKE), the Horizontal Pod and Cluster Autoscaler can scale things, provided they're configured. ⚙️
That's not a default setting, it needs the right configuration, and database/backend must not become the bottleneck either.
How ready is your infrastructure for the next spike?
-
The hailstorm last week pushed several Swiss insurance claim portals to their limits. ⛈️
Portals were temporarily overloaded or unreachable.
With Managed Kubernetes (NKE), the Horizontal Pod and Cluster Autoscaler can scale things, provided they're configured. ⚙️
That's not a default setting, it needs the right configuration, and database/backend must not become the bottleneck either.
How ready is your infrastructure for the next spike?
-
The hailstorm last week pushed several Swiss insurance claim portals to their limits. ⛈️
Portals were temporarily overloaded or unreachable.
With Managed Kubernetes (NKE), the Horizontal Pod and Cluster Autoscaler can scale things, provided they're configured. ⚙️
That's not a default setting, it needs the right configuration, and database/backend must not become the bottleneck either.
How ready is your infrastructure for the next spike?
-
The hailstorm last week pushed several Swiss insurance claim portals to their limits. ⛈️
Portals were temporarily overloaded or unreachable.
With Managed Kubernetes (NKE), the Horizontal Pod and Cluster Autoscaler can scale things, provided they're configured. ⚙️
That's not a default setting, it needs the right configuration, and database/backend must not become the bottleneck either.
How ready is your infrastructure for the next spike?
-
Автоскейлинг StarRocks в Kubernetes: как я довел его до предела
Классическая проблема аналитических систем: кластер СУБД сайзится под пик, а 28 дней в месяц он задействован чуть больше чем наполовину. StarRocks (shared-data) и автоскейл Kubernetes убирают этот избыток. Compute добавляется под нагрузку и сворачивается на спаде. Внутри легкая пятничная статья: как это работает и где у эластичности потолок.
https://habr.com/ru/articles/1041278/
#starrocks #olap #kubernetes #autoscaling #selena #субд #mpp