还是别回头了吧, 你既不拥有过去, 也曾未到达未来.
Entertainment
Young Sheldon S1
What a nerd, right? 相比于原著的大爆炸理论, 个人觉得更加好看一点; 至少, 我对 lenerd 或者说其他人之间所演绎的“感情戏”, or just say adult sex life, 一点都不感兴趣.
家庭, 成长这些主题则可以推广到范围更加庞大的受众群, 也许能给有类似情景下的“当事者”一些启发.
Learning
I was preparing for the SG exam, so did a bunch of exam simulation.
Life
- 三连休的结束, 出门爬山, 也算是体验了一把“乡下的繁忙”景象
- 新宿歌舞伎町, 吃了来日之后第一份海底捞, 个人感觉整体就是照搬过来的, 服务模式等等; 味道也还算 OK, 价格也没有超出预算太多
- 平日的早晚以及周末的时间, 个人评价还是浪费了其中的绝大部分, 一方面是目标明确了, 但个人的认知没有跟上; 一方面是, 可能还是“承认欲求”在作祟吧, 孤独感与社群融合的纠结
- 萨利亚还不错, 这周去了两趟, 主要去吃鸡翅, 还是相当划算的; 但, 感觉还是有点超出预算, 到底是怎么了呢 (我的生活)
Collectibles
learn
- https://www.uber.com/en-JP/blog/better-load-balancing-real-time-dynamic-subsetting
- how to design load balancing system under a massive service mesh setting
- Subset for the rescue, but the metrics are critical for the fairly distributing
- three key factors in acticle
- traffic assignment - the percentage of traffic to send to each pool
- load - the amount of traffic it’s sending to a target service
- aggregate load - the overall traffic a target service is receiving
- $desiredSubsetSize = numberOfTaskInPool * load/(aggregateLoad*assignment) * constant$
- https://www.uber.com/en-IN/blog/load-balancing-handling-heterogeneous-hardware
- on-host agent, load balancing strategy (inter-cluster, intra-cluster)
- what problem to consider when we talk about metric
- any stat collection was deplayed
- the service was restarted
- instead of attaching the load meta-data to each of the responses, using a central component to collect and distribute the data.
- reponse-header-based
share
- breath-taking views
- Milano https://youtu.be/qS-r-vJpnvo
- 大兴安岭 https://youtu.be/BlsTrHqlNMo
- Svalbard part1 https://youtu.be/mt8UkFYfSBk
- Svalbard part2 https://youtu.be/qYTH1SWkW_s