投稿時間:2023-07-17 05:14:42 RSSフィード2023-07-17 05:00 分まとめ(18件)

カテゴリー等	サイト名等	記事タイトル・トレンドワード等	リンクURL	頻出ワード・要約等/検索ボリューム	登録日
Ruby	Railsタグが付けられた新着投稿 - Qiita	ngrokを用いたローカル開発環境の公開方法	https://qiita.com/kzsiaaidd/items/9b6e67c0d8a4fe25277b	linemessagingapi	2023-07-17 04:35:29
海外TECH	MakeUseOf	How to Fix Microsoft Outlook's “Something Went Wrong” Error on Windows	https://www.makeuseof.com/microsoft-outloos-something-went-wrong-error-windows/	How to Fix Microsoft Outlook x s “Something Went Wrong Error on WindowsSometimes error messages aren t great at telling you why an app crashed Here s how to fix the annoyingly vague quot something went wrong quot Outlook error	2023-07-16 19:16:18
海外TECH	MakeUseOf	Is WhatsApp Safe? 6 Scams, Threats, and Security Risks to Know About	https://www.makeuseof.com/tag/4-security-threats-whatsapp-users-need-know/	Is WhatsApp Safe Scams Threats and Security Risks to Know AboutWhatsApp security is a challenge making it a frequent target of scammers and hackers Is WhatsApp safe Here s what you need to know	2023-07-16 19:06:19
海外TECH	DEV Community	Building Serverless Applications with AWS – Compute	https://dev.to/aws-builders/building-serverless-applications-with-aws-compute-2704	Building Serverless Applications with AWS ComputeBuilding Serverless applications can feel a bit overwhelming when you are first getting started Sure Event Driven Systems have been around for many years but this notion of using managed services to assemble solutions vs a more traditional plugin style architecture might throw you for a loop Continuing in the series of Building Serverless Applications with AWS let s have a look at the Compute aspect Series Topics Data Storage Choices Building the Application Fargate Containers vs Lambda Handling Events Exposing the API if there is one Securing it all including the API Debugging and Troubleshooting in Production Building Serverless Applications ComputeOne could make the argument that compute portion of the Serverless stack might be the most important I m not sure that I agree with that however it is a critical component as it s where your code runs I mostly believe that your app is going to have code running somewhere regardless of whether it s Serverless or Serverfull and that all of the supporting services are really what makes this paradigm shine Moving on from the pontification I see compute being divided into two different runtimes when I m architecting Serverless solutions Those two are AWS Fargate and AWS Lambda I d first like to define what category of compute these two are in my mind and then start to break apart the differences and the nuance of putting them into your build AWS FargateAWS Fargate is a serverless pay as you go compute engine that lets you focus on building applications without managing servers AWS Fargate is compatible with both Amazon Elastic Container Service ECS and Amazon Elastic Kubernetes Service EKS AWS And to further illustrate here is their infographicThink about it this way You build a container image Docker and then based on that image s memory and CPU requirements you ask Fargate to create a Task with those given parameters AWS LambdaCompare that with AWS Lambda which is described by AWS like this AWS Lambda is a serverless event driven compute service that lets you run code for virtually any type of application or backend service without provisioning or managing servers You can trigger Lambda from over AWS services and software as a service SaaS applications and only pay for what you use AWSThe way I think about when to use Lambdas is largely by what is going to trigger them and if the demand is consistent or possibly bursty What I mean by this is that if I have a constant demand I might reach for Fargate because I can scale predictably But this is really only a might and generally boils down to who is doing the building and the supporting I ll explain more as I get into the article The Compute CriteriaSo if we want to look first at perhaps some hard criteria to understand one vs the other these are the questions that I might ask Does your code run in bursts or perhaps not consistently throughout the day Does the idea of only paying for resources used to appeal to you Are you connecting to other Serverless services and using those things as Event Triggers Do you have a lot of passion for containers or prefer to just write your code and let AWS take care of the rest Side note sure you can deploy custom images into the Lambda runtime but controversially I say why If you answered yes to those questions and maybe no to the last I would tend to think that you are going to be looking to deploy in Lambda But let s look at the converse of that why you might look at Fargate Are you bringing legacy code to the cloud that s currently already running on prem or maybe even in EC Is your workload more consistent and even balanced Can you perhaps shut things off after hours or will you have steady demand after hours Is your scaling predictable or are you comfortable scaling up to meet demand with alarms and triggers to provision new infrastructure Do you have a longer running task that instead of using Step Functions you d feel better running in a host Lambda will drop you at minutes so this might matter Answering yes to the above lends me to thinking about Fargate If you are looking for building consistent workloads with more Opensource and non AWS frameworks then go with Fargate It s going to give you that EC but still Serverless feel Compute SummaryThe real truth here is that you can make some awesome and valuable applications with either of these services I tend to use Lambda over Fargate but that s due more to the way I like to code and ship than Lambda being any better than Fargate I know other peers that like Containers so much more than Functions and they choose Fargate all day It s a matter of what fits what you are trying to accomplish So for the balance of this article I want to address how to operate these two services to better give you a view of how to choose between them Working with these Compute ServicesWhat does this mean though for you as a developer or architect Again let me state that you can have success with either of these services in all the same use cases So unless you ve got something like the gt minutes greater than rule or you have a really large binary you can run both of these just fine I m going to break each of these down into the following categories Deploying your code Configuring common parameters Cost Overall Developer experience GotchasPlease remember these are my takes from my experiences It s not exhaustive and I ll never profess to be an expert at anything and especially something that is so open ended as Serverless But in the end as a new developer to Serverless I hope you find them helpful Deploying your CodeI m going to start with Lambda and then pivot into Fargate as I work through these elements For a Lambda you have two approaches for getting your code to where it can be executed Path A is to create a container image where you can use SAM or using the Docker CLI to build your image from one of the AWS bases Path B is to create a Lambda Deployment from a zip file that is an archive of sorts Each archive might look a little different depending upon your language of choice There is a list of support runtimes including custom ones that you can leverage Whichever path you follow A or B the code needs to be ready to handle the event that you are going to be receiving What makes Lambda very different from running Fargate is that the entry point to your function is going to be a handler with a specific event payload it expects Root level it could be JSON or it could be something strongly typed Again it really will depend upon your language This is a stark contrast to what you ll see in Fargate below With Fargate things are a little more standard from a non Serverless perspective You need a container image that will execute your code This code can then be whatever you wish For instance let s say you have a Java Spring Boot web server or a Golang Gin web server You ll need a main and then that code will fire off your web server It s going to feel just like if you were running in EC or on your laptop What I like about Fargate is that if you are coming from a more traditional background getting into Serverless is much more straightforward This is especially true if you are coming from previous Docker experience I ve seen many developers start here before they ever deploy their first Lambda function Configuring Common ParametersUnderstanding how your Serverless service works and how it can be tuned can be one of the biggest things that take time and experimentation As a developer it s as important as learning the API of any other dependency in your stack but sometimes learning it takes running it Hopefully the below will help you gain some insight into the things that will initially matter There is a great deal more than what I m about to share but for starters these are the things that are most important in my eyes when one s learning these compute platforms Lambda ParametersFor Lambda the things that matter the most to me are this and in no order of importance Memory allocated This tells Lambda how much memory is assigned to your function and also proportionately assigns vCPU There s no setting for CPU but you ll see there will be a sweet spot for how much memory balances against performance Environment variables Seems simple right But having the right variables configured makes your Lambda more adaptable so that you can alter behavior without changing code VPC Settings I m going to be honest I rarely ever use this But when you need to use something inside your VPC you ll need to specify the VPC and the Subnets that you want your Lambda attached to IAM Policies Your Lambda will get its permissions from the policies that are attached to it When you look at using the SDK for calling any other services you won t need an ACCESS KEY ID or SECRET ACCESS KEY as the SDK will pick those up from your Lambda when it runs So pay attention to what you grant in there and only grant what you need Concurrency By default your account will only allow so many Lambda invocations at one time Additionally you might need to restrict the number of versions of this code running at any one time Setting the max concurrency on your Lambda can be important Last and it s a big last Learn the configurations for the Events that will trigger your Lambda For SQS how many messages per poll For Kinesis what s the batch size and what happens on failures So many things to learn here Fargate ParametersWhen working with Fargate I can t write up anything much better than what s already done in the AWS documentation But what I will do is point out what I think is important when you start So those key things are this vCPU This means just what you think it does How much many CPUs do you want to allocate to your container Memory Also super straightforward You need to balance vCPU with Memory as there are combinations that you can t create But in general you should find this to be what you expect IAM Policies Your Fargate Task will get its permissions from the policies that are attached to it When you look at using the SDK for calling any other services you won t need an ACCESS KEY ID or SECRET ACCESS KEY as the SDK will pick those up from your Task when it runs So pay attention to what you grant in there and only grant what you need Docker Settings Which ports are you opening and which port is the internal one listening to They might be different What s the HealthCheck endpoint Compute CostNow here s where there is a good bit of difference and something worth thinking about Lambda CostFor Lambda first here is what AWS says Lambda counts a request each time it starts executing in response to an event notification trigger such as from Amazon Simple Notification Service SNS or Amazon EventBridge or an invoke call such as from Amazon API Gateway or via the AWS SDK including test invokes from the AWS Console Duration is calculated from the time your code begins executing until it returns or otherwise terminates rounded up to the nearest ms The price depends on the amount of memory you allocate to your function In the AWS Lambda resource model you choose the amount of memory you want for your function and are allocated proportional CPU power and other resources An increase in memory size triggers an equivalent increase in CPU available to your function AWSWhat the above means is that you are charged for the size and duration of your function where size is measured in Memory Allocated Take for instance a simple Queue event handler that only runs a handful of times a day It really makes very little sense to have a container in Fargate running for something like this Of course you could spin it up spin it down and whatnot but why take that overhead Additionally you might have a Lambda that runs a million times a day I could make the argument that the Fargate task would be cheaper if the request overhead makes sense but then ask what if the container doesn t need to be up overnight It might be a mix of Fargate and then turning it off Or perhaps even further you don t want to worry about it and you are willing to pay the overhead to run Lambda all the time That s your choice as well With anything unless you are racking your equipment perfectly sizing and going bare metal one can always argue cost I wrote about what I think is the true cost of Serverless so might be worth taking a read Fargate CostPer AWS here s how Fargate pricing works AWS Fargate pricing is calculated based on the vCPU memory Operating Systems CPU Architecture and storage resources used from the time you start to download your container image until the Amazon ECS Task or Amazon EKS Pod terminates rounded up to the nearest second AWSPricing is based on requested vCPU memory Operating Systems CPU Architecture and storage resources for the Task or Pod The five dimensions are independently configurable AWSBottom line you are going to get charged for the architecture storage compute and memory that you need From there it s easy You don t need to worry about where the container is deployed or even how just that you have an IP address that is in front of your container that is running your code and exposing the ports you ve asked for with the parameter settings you specified You then pay for what you use while it s running Not events Just time running You could handle millions of events an hour with a static price by the number of nodes times the chargeable parameters Additionally you could pay that same price if you do no traffic over that same period It s for that reason that even though it s Serverless you are paying for time running and not events handled So use that to your advantage Overall Developer experience GotchasOK so this is super subjective so I m not going to talk about which one is better than the other I m already on the record that I love Lambdas and tend to use them more But when I ve got very consistent workloads and want maybe a little more control over my frameworks I go Fargate The nice thing about Fargate too is that you can leverage ECS or EKS as the container management platform from AWS I won t begin to step into that battle but for me and probably you unless you are a Kubernetes ninja is to just start with ECS Lambda ExperiencesThe things that I think you need to be careful about are as follows when it comes to Lambda Pay attention to where your logs are going You are going to need to debug and logs matter Use a JSON formatter where possible Watch exceptions throttles and concurrency Remember you are looking event over event Very much like a request Those metrics are going to matter as you get better at using Lambda Seriously learn to watch these like a hawk Downstream needs and concurrency Again with the concurrency I ve seen Lambdas flood downstream resources because AWS will keep spinning up more Lambda nodes regardless of what other systems require If you have a pool of DB connections and your DB can deal with connections but each of your Lambdas opens connections Guess what Lambda invocations will render your DB useless WATCH THIS If you ve got shared code look at Layers or Extensions These are powerful and I think under used Function timeout How long will your code be allowed to execute You have a max of minutes but let s say you forget to deal with failure If so your Lambda will hang on for the timeout you set thus racking of the bill for code that s wasted Embrace testing in the cloud I know I love local development but I m slowly coming around It ll make your life easier Fargate ExperiencesMy opinions of working with Fargate are so positive and straightforward that my things to learn from are pretty simple If your container doesn t need to be up shut it down It ll save you cost If you need more capacity burst sooner Your service might take seconds to become healthy You d hate to have users experience that unavailable moment because you scaled too late You might not need as much memory and CPU as you think Play with this Make your HealthCheck robust This is just good container management but I felt I needed to say it You are going to get charged to pull your image So invest in small images I cannot say this enough If you are looking to migrate to Serverless taking Spring Boot ASP NET Core a Golang web server and putting in Fargate is easy Especially considering that you might already have the image defined Then it s even that much easier Pretty simple list honestly I love Fargate I don t use it nearly as much as I think I should but it s a fantastic service and something you need to be using Wrapping UpAlright so that was words on Compute I didn t intend to do that much but as you can see from the above if you are still with me there s a lot to process Pun intended This to me is one of those things where you really can t go wrong AWS has done an amazing job giving Serverless developers and architects two world class compute platforms If you want per invocation control out of the box events pay as you go pricing then Lambda is your choice If you want a more familiar Docker experience with higher level control than what you d have with fine grained EC then Fargate is your huckleberry I also think that you will mix and match these as you see fit and as you learn more and run more When building in Serverless things are often small so you aren t locked in Be my guest to experiment and see what works best for your environment Most of my guidance above comes from s of millions of executions on both platforms and learning about how they operate and how I can bring the best value to my end users I sincerely hope that you ve gained some knowledge that ll help you do the same thing Until the next in the series Happy Building	2023-07-16 19:08:33
ニュース	BBC News - Home	Carlos Alcaraz beats Novak Djokovic to win Wimbledon men's title	https://www.bbc.co.uk/sport/tennis/66212814?at_medium=RSS&at_campaign=KARANGA	Carlos Alcaraz beats Novak Djokovic to win Wimbledon men x s titleCarlos Alcaraz wins the Wimbledon men s singles title for the first time by ending Novak Djokovic s recent dominance with a stunning victory	2023-07-16 19:08:51
ニュース	BBC News - Home	Iran's morality police to resume headscarf patrols	https://www.bbc.co.uk/news/world-middle-east-66218318?at_medium=RSS&at_campaign=KARANGA	dress	2023-07-16 19:10:04
ニュース	BBC News - Home	Women's Ashes: Australia retain Ashes despite Nat Sciver-Brunt's unbeaten 111 - highlights	https://www.bbc.co.uk/sport/av/cricket/66218305?at_medium=RSS&at_campaign=KARANGA	Women x s Ashes Australia retain Ashes despite Nat Sciver Brunt x s unbeaten highlightsWatch the highlights as Australia retained the Women s Ashes with a dramatic three run victory over England who so nearly pulled off a remarkable win through Nat Sciver Brunt s masterful unbeaten	2023-07-16 19:20:19
ビジネス	ダイヤモンド・オンライン - 新着記事	タワマン節税はもう“使えない”！新ルールとタワマン「増税リスク」を徹底検証 - タワマン節税 緊急事態	https://diamond.jp/articles/-/326058	マンション評価改正を取り巻く最新状況を解説するとともに、自身の物件のリスクを正しく判断し対策につなげるための、マンション増税の影響が見込まれる物件ランキングを紹介する。	2023-07-17 04:50:00
ビジネス	ダイヤモンド・オンライン - 新着記事	知られざるエリートサークル「大学弁論部」を大解剖！100年超の秘史と政財官に根付く人脈 - 知られざるエリート人脈 大学弁論部の正体	https://diamond.jp/articles/-/326118	知られざるエリートサークル「大学弁論部」を大解剖年超の秘史と政財官に根付く人脈知られざるエリート人脈大学弁論部の正体明治期の自由民権運動とともに発展してきた大学弁論部は、「弁論」を研さんする場として長い歴史を持つ。	2023-07-17 04:45:00
ビジネス	ダイヤモンド・オンライン - 新着記事	投資を始めると会社で出世する決定的理由 - 10万円から始める! 小型株集中投資で１億円 【１問１答】株ドリル	https://diamond.jp/articles/-/324330	投資を始めると会社で出世する決定的理由万円から始める小型株集中投資で億円【問答】株ドリル【大好評シリーズ万部突破】東京理科大学の大学生だったとき、夏休みの暇つぶしで突如「そうだ、投資をしよう」と思い立った。	2023-07-17 04:20:00
ビジネス	ダイヤモンド・オンライン - 新着記事	【全米屈指のデータサイエンティストが教える】今日から貯金が楽しくなるアドバイス・ベスト1 - JUST KEEP BUYING	https://diamond.jp/articles/-/325279	【全米屈指のデータサイエンティストが教える】今日から貯金が楽しくなるアドバイス・ベストJUSTKEEPBUYING【発売たちまち大重版】全世界万部突破『サイコロジー・オブ・マネー』著者モーガン・ハウセルが「絶対読むべき一冊」と絶賛。	2023-07-17 04:17:00
ビジネス	ダイヤモンド・オンライン - 新着記事	「不幸の宝くじ」である保険は、損をすることに意味がある金融商品 - シンプルで合理的な人生設計	https://diamond.jp/articles/-/326025	人生設計	2023-07-17 04:14:00
ビジネス	ダイヤモンド・オンライン - 新着記事	国士舘大学のキャンパス環境はどんな雰囲気？【各キャンパス紹介付き】 - 大学図鑑！2024 有名大学82校のすべてがわかる！	https://diamond.jp/articles/-/326158		2023-07-17 04:11:00
ビジネス	ダイヤモンド・オンライン - 新着記事	職場にいる「本当に深く考えている人」と「考えていそうで浅い人」の決定的な差とは - 1秒で答えをつくる力 お笑い芸人が学ぶ「切り返し」のプロになる48の技術	https://diamond.jp/articles/-/326145		2023-07-17 04:08:00
ビジネス	ダイヤモンド・オンライン - 新着記事	経営者が「社員に任せて育成する」のを今すぐやめるべき理由（前編） - 新装版　売上2億円の会社を10億円にする方法	https://diamond.jp/articles/-/325756	経営者が「社員に任せて育成する」のを今すぐやめるべき理由前編新装版売上億円の会社を億円にする方法コツコツと業績を伸ばしてきた経営者が直面する「売上の壁」。	2023-07-17 04:05:00
ビジネス	ダイヤモンド・オンライン - 新着記事	【精神科医が教える】 他人の悪口で盛り上がったとき、見失っていること - 精神科医Tomyが教える 40代を後悔せず生きる言葉	https://diamond.jp/articles/-/325865	【精神科医が教える】他人の悪口で盛り上がったとき、見失っていること精神科医Tomyが教える代を後悔せず生きる言葉【大好評シリーズ万部突破】誰しも悩みや不安は尽きない。	2023-07-17 04:02:00
ビジネス	不景気.com	旭化成がタイヤ補強材の生産から撤退、宮崎の工場を閉鎖 - 不景気com	https://www.fukeiki.com/2023/07/asahi-kasei-code-close.html	化学メーカー	2023-07-16 19:07:16
ビジネス	東洋経済オンライン	JR東日本の技術｢東海の駅｣で実用化､その意味は? 系列ベンチャー企業が開発の無人決済システム | 経営 | 東洋経済オンライン	https://toyokeizai.net/articles/-/686624?utm_source=rss&utm_medium=http&utm_campaign=link_back	東洋経済オンライン	2023-07-17 04:30:00