<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Robin Sishodia</title>
    <description>The latest articles on DEV Community by Robin Sishodia (@robinsishodia).</description>
    <link>https://dev.to/robinsishodia</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4213306%2F85d72d36-dc3f-4a0e-9f8a-e02f7558a2c3.jpg</url>
      <title>DEV Community: Robin Sishodia</title>
      <link>https://dev.to/robinsishodia</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/robinsishodia"/>
    <language>en</language>
    <item>
      <title>AWS Secrets Manager and Parameter Store: Keep Secrets Out of Your Code</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Thu, 19 Nov 2026 17:08:51 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-secrets-manager-and-parameter-store-keep-secrets-out-of-your-code-349p</link>
      <guid>https://dev.to/robinsishodia/aws-secrets-manager-and-parameter-store-keep-secrets-out-of-your-code-349p</guid>
      <description>&lt;p&gt;Systems Manager gave me a way into instances without SSH. The next question is where an application should keep its &lt;strong&gt;database password and settings&lt;/strong&gt; so they never end up in code. AWS has two good answers: &lt;strong&gt;Secrets Manager&lt;/strong&gt; and &lt;strong&gt;Parameter Store&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  The two services
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;AWS Secrets Manager&lt;/strong&gt; stores secrets such as database credentials and API keys, encrypts them with &lt;strong&gt;AWS KMS&lt;/strong&gt;, and can &lt;strong&gt;rotate&lt;/strong&gt; them automatically.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Systems Manager Parameter Store&lt;/strong&gt; stores configuration values and secrets in a &lt;strong&gt;path-style hierarchy&lt;/strong&gt;, such as &lt;code&gt;/myapp/prod/db-host&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Both are accessed with an API call and controlled with &lt;strong&gt;IAM permissions&lt;/strong&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  How they compare
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;&lt;/th&gt;
&lt;th&gt;Secrets Manager&lt;/th&gt;
&lt;th&gt;Parameter Store&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Best for&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Credentials that need rotation&lt;/td&gt;
&lt;td&gt;App settings and simple secrets&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Rotation&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Built in, using Lambda&lt;/td&gt;
&lt;td&gt;Not built in&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Encryption&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Always, with KMS&lt;/td&gt;
&lt;td&gt;Only for the &lt;strong&gt;SecureString&lt;/strong&gt; type&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Cost&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Billed per secret per month, plus API calls&lt;/td&gt;
&lt;td&gt;Standard parameters are free; Advanced parameters cost extra&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Size&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Larger values&lt;/td&gt;
&lt;td&gt;Standard parameters are up to 4 KB&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Check the current pricing page before you rely on these numbers.&lt;/p&gt;

&lt;h2&gt;
  
  
  Parameter types
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;String&lt;/strong&gt; for plain text such as a hostname&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;StringList&lt;/strong&gt; for comma-separated values&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;SecureString&lt;/strong&gt; for values encrypted with KMS&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Access with IAM
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;The caller needs permission such as &lt;code&gt;secretsmanager:GetSecretValue&lt;/code&gt; or &lt;code&gt;ssm:GetParameter&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;For encrypted values it also needs permission to use the KMS key, such as &lt;code&gt;kms:Decrypt&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Give an application role only the paths or secrets it needs.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;p&gt;Use a fake value only. Never put a real password in a practice exercise.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Store a plain parameter: &lt;code&gt;aws ssm put-parameter --name /demo/app/db-host --value mydb.example.internal --type String&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;Store an encrypted one: &lt;code&gt;aws ssm put-parameter --name /demo/app/db-password --value FakePassw0rd --type SecureString&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;Read it back: &lt;code&gt;aws ssm get-parameter --name /demo/app/db-password --with-decryption&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;List by path: &lt;code&gt;aws ssm get-parameters-by-path --path /demo/app --with-decryption&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;Create a secret: &lt;code&gt;aws secretsmanager create-secret --name demo/db-credentials --secret-string '{"username":"admin","password":"FakePassw0rd"}'&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;Read it: &lt;code&gt;aws secretsmanager get-secret-value --secret-id demo/db-credentials&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;In the console, open the secret and look at the &lt;strong&gt;Rotation&lt;/strong&gt; settings.&lt;/li&gt;
&lt;li&gt;Clean up: &lt;code&gt;aws ssm delete-parameter --name /demo/app/db-host&lt;/code&gt;, the same for &lt;code&gt;/demo/app/db-password&lt;/code&gt;, and &lt;code&gt;aws secretsmanager delete-secret --secret-id demo/db-credentials --force-delete-without-recovery&lt;/code&gt;
&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Hardcoding passwords in code, in environment files or in a repository.&lt;/li&gt;
&lt;li&gt;Storing a secret as a plain &lt;strong&gt;String&lt;/strong&gt; parameter.&lt;/li&gt;
&lt;li&gt;Giving a role access to every secret instead of the ones it needs.&lt;/li&gt;
&lt;li&gt;Printing secret values in logs or in screenshots.&lt;/li&gt;
&lt;li&gt;Forgetting to delete practice secrets, which keep costing money.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Keep secrets out of code, and fetch them at runtime.&lt;/li&gt;
&lt;li&gt;Use Secrets Manager when you need rotation, and Parameter Store for settings and simple secrets.&lt;/li&gt;
&lt;li&gt;Use SecureString for anything sensitive in Parameter Store.&lt;/li&gt;
&lt;li&gt;Limit access by path or secret name, and clean up after practice.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at AWS KMS and how the encryption keys behind these services work.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS Systems Manager: Manage Your Instances Without Opening SSH</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Wed, 18 Nov 2026 14:27:14 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-systems-manager-manage-your-instances-without-opening-ssh-1ljo</link>
      <guid>https://dev.to/robinsishodia/aws-systems-manager-manage-your-instances-without-opening-ssh-1ljo</guid>
      <description>&lt;p&gt;Config shows what changed and whether it follows the rules. The next question is how to actually &lt;strong&gt;manage&lt;/strong&gt; the instances without opening SSH to the world. That is where &lt;strong&gt;AWS Systems Manager&lt;/strong&gt; comes in.&lt;/p&gt;

&lt;h2&gt;
  
  
  What it does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Systems Manager is a set of tools to &lt;strong&gt;view and manage&lt;/strong&gt; your AWS resources and nodes from one place.&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;managed node&lt;/strong&gt; is an instance (or server) that runs the &lt;strong&gt;SSM Agent&lt;/strong&gt; and can reach Systems Manager.&lt;/li&gt;
&lt;li&gt;You can connect, run commands and patch &lt;strong&gt;without opening inbound port 22&lt;/strong&gt; and without key pairs.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The tools to know first
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Tool&lt;/th&gt;
&lt;th&gt;What it does&lt;/th&gt;
&lt;th&gt;Example use&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Session Manager&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Opens a shell from the console or CLI&lt;/td&gt;
&lt;td&gt;Log in with no SSH and no bastion host&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Run Command&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Runs a command on one or many nodes&lt;/td&gt;
&lt;td&gt;Run &lt;code&gt;uptime&lt;/code&gt; on a group of instances&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Patch Manager&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Scans and installs patches&lt;/td&gt;
&lt;td&gt;Patch on a schedule&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Parameter Store&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Stores configuration values and secrets&lt;/td&gt;
&lt;td&gt;Keep a database name or a secure string&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Automation&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Runs runbooks for common tasks&lt;/td&gt;
&lt;td&gt;Stop an instance, create an AMI&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What an instance needs to become a managed node
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;The &lt;strong&gt;SSM Agent&lt;/strong&gt; installed and running (it comes pre-installed on many Amazon-provided AMIs, including Amazon Linux 2023).&lt;/li&gt;
&lt;li&gt;An &lt;strong&gt;instance profile&lt;/strong&gt; with an IAM role, for example the AWS managed policy &lt;code&gt;AmazonSSMManagedInstanceCore&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;A network path to the Systems Manager endpoints, such as a public route or &lt;strong&gt;VPC endpoints&lt;/strong&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;IAM&lt;/strong&gt;, create a role for &lt;strong&gt;EC2&lt;/strong&gt;, and attach &lt;code&gt;AmazonSSMManagedInstanceCore&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Launch a small &lt;strong&gt;Amazon Linux 2023&lt;/strong&gt; instance and choose that role as the &lt;strong&gt;IAM instance profile&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;In the security group, &lt;strong&gt;remove all inbound rules&lt;/strong&gt;. You will not use SSH.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Systems Manager&lt;/strong&gt;, then &lt;strong&gt;Fleet Manager&lt;/strong&gt; or &lt;strong&gt;Session Manager&lt;/strong&gt;, and wait until the instance shows as a managed node.&lt;/li&gt;
&lt;li&gt;In &lt;strong&gt;Session Manager&lt;/strong&gt;, choose &lt;strong&gt;Start session&lt;/strong&gt;, pick the instance, and run &lt;code&gt;whoami&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Run Command&lt;/strong&gt;, choose the document &lt;code&gt;AWS-RunShellScript&lt;/code&gt;, enter &lt;code&gt;uptime&lt;/code&gt; and run it on the instance.&lt;/li&gt;
&lt;li&gt;Check from the CLI: &lt;code&gt;aws ssm describe-instance-information&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;When you finish practising, terminate the instance and delete the role.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Forgetting to attach the instance profile, so the node never appears.&lt;/li&gt;
&lt;li&gt;Using an instance with no route to the Systems Manager endpoints.&lt;/li&gt;
&lt;li&gt;Leaving port 22 open even though Session Manager makes it unnecessary.&lt;/li&gt;
&lt;li&gt;Giving the instance role more permissions than it needs.&lt;/li&gt;
&lt;li&gt;Leaving practice instances running.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Systems Manager manages nodes without SSH or key pairs.&lt;/li&gt;
&lt;li&gt;An instance needs the agent, an instance profile and network access.&lt;/li&gt;
&lt;li&gt;Session Manager is for access, Run Command for tasks, Patch Manager for updates.&lt;/li&gt;
&lt;li&gt;Clean up practice instances when you finish.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at AWS Secrets Manager and Parameter Store, and how to keep secrets out of your code.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS Config: Track Resource Changes and Check Compliance</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Tue, 17 Nov 2026 16:10:26 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-config-track-resource-changes-and-check-compliance-1h46</link>
      <guid>https://dev.to/robinsishodia/aws-config-track-resource-changes-and-check-compliance-1h46</guid>
      <description>&lt;p&gt;CloudTrail tells me who made a change. The next question is different: &lt;strong&gt;what did the resource look like before, and does it follow my rules?&lt;/strong&gt; That is the job of &lt;strong&gt;AWS Config&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What AWS Config does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;It &lt;strong&gt;records the configuration&lt;/strong&gt; of supported AWS resources and keeps a history of how each one changed.&lt;/li&gt;
&lt;li&gt;It checks those configurations against &lt;strong&gt;rules&lt;/strong&gt; and reports each resource as &lt;strong&gt;compliant&lt;/strong&gt; or &lt;strong&gt;non-compliant&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;It works &lt;strong&gt;per Region&lt;/strong&gt;, so you set it up in each Region you use.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  CloudTrail and Config together
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Question&lt;/th&gt;
&lt;th&gt;Service&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Who made this API call, and when?&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;CloudTrail&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;What did this resource look like last week?&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;Config&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Is this resource following our rule?&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;Config&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  The building blocks
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Term&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Configuration recorder&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Records resource settings&lt;/td&gt;
&lt;td&gt;Choose all resources or specific types&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Configuration item&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A snapshot of one resource at a point in time&lt;/td&gt;
&lt;td&gt;Builds the resource timeline&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Config rule&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A check on a resource setting&lt;/td&gt;
&lt;td&gt;AWS managed rules or your own&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Compliance status&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Result of a rule&lt;/td&gt;
&lt;td&gt;COMPLIANT or NON_COMPLIANT&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Remediation&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;An action to fix a non-compliant resource&lt;/td&gt;
&lt;td&gt;Can run an SSM Automation document&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Examples of managed rules
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;s3-bucket-public-read-prohibited&lt;/code&gt; flags buckets that allow public reads.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;restricted-ssh&lt;/code&gt; flags security groups that allow SSH from anywhere.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;encrypted-volumes&lt;/code&gt; flags EBS volumes that are not encrypted.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;AWS Config&lt;/strong&gt; and choose &lt;strong&gt;Set up&lt;/strong&gt; (or &lt;strong&gt;Get started&lt;/strong&gt;).&lt;/li&gt;
&lt;li&gt;Record &lt;strong&gt;specific resource types&lt;/strong&gt; such as S3 buckets and security groups, to keep it small.&lt;/li&gt;
&lt;li&gt;Let it create an S3 bucket for the delivery channel and the service-linked role.&lt;/li&gt;
&lt;li&gt;Add the managed rule &lt;code&gt;s3-bucket-public-read-prohibited&lt;/code&gt;, then add &lt;code&gt;restricted-ssh&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Wait for the first evaluation, then open &lt;strong&gt;Rules&lt;/strong&gt; and read the compliance status.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Resources&lt;/strong&gt;, pick a security group and read its &lt;strong&gt;resource timeline&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Check from the CLI: &lt;code&gt;aws configservice describe-compliance-by-config-rule&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;When you finish practising, stop the recorder: &lt;code&gt;aws configservice stop-configuration-recorder --configuration-recorder-name default&lt;/code&gt;
&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Setting up Config in one Region and expecting it everywhere.&lt;/li&gt;
&lt;li&gt;Recording every resource type for a small practice account without checking the cost.&lt;/li&gt;
&lt;li&gt;Adding rules but never looking at the non-compliant list.&lt;/li&gt;
&lt;li&gt;Forgetting to stop the recorder after practice.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Config records how resources change and checks them against rules.&lt;/li&gt;
&lt;li&gt;CloudTrail answers who, Config answers what and whether it is compliant.&lt;/li&gt;
&lt;li&gt;Start with a few managed rules and review the results.&lt;/li&gt;
&lt;li&gt;Watch the cost, because it depends on what you record and evaluate.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at AWS Systems Manager and how it manages instances without opening SSH.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS CloudTrail: See Who Did What in Your AWS Account</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Mon, 16 Nov 2026 14:16:57 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-cloudtrail-see-who-did-what-in-your-aws-account-59ai</link>
      <guid>https://dev.to/robinsishodia/aws-cloudtrail-see-who-did-what-in-your-aws-account-59ai</guid>
      <description>&lt;p&gt;Last time I set up an alarm to be told when something goes wrong. The next question is different: &lt;strong&gt;who did what in my account?&lt;/strong&gt; When a setting changes or a resource disappears, the answer comes from &lt;strong&gt;AWS CloudTrail&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What CloudTrail does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;It records &lt;strong&gt;API activity&lt;/strong&gt; in your account: who made the call, what they did, when, and from where.&lt;/li&gt;
&lt;li&gt;That includes actions in the console, the CLI, SDKs and other AWS services.&lt;/li&gt;
&lt;li&gt;It is the first place to look after a surprise change or a security question.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The building blocks
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Term&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Event&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;One recorded API call&lt;/td&gt;
&lt;td&gt;Shows user, action, time, source IP&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Event history&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A searchable view of recent events&lt;/td&gt;
&lt;td&gt;Management events, last 90 days, per Region&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Trail&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A setting that delivers events to an S3 bucket&lt;/td&gt;
&lt;td&gt;Keeps history longer than 90 days&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Management events&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Actions on resources&lt;/td&gt;
&lt;td&gt;Examples: CreateTopic, RunInstances, DeleteBucket&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Data events&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Actions on the data inside resources&lt;/td&gt;
&lt;td&gt;Examples: S3 object reads, Lambda invokes. Off by default&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Event history or a trail?
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Event history&lt;/strong&gt; is on by default, needs no setup and covers the last 90 days of management events.&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;trail&lt;/strong&gt; stores events in S3 so you can keep them, search them later and share them across the organization.&lt;/li&gt;
&lt;li&gt;A trail can apply to &lt;strong&gt;all Regions&lt;/strong&gt;, so activity in a Region you never use still gets recorded.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: find who did something
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;CloudTrail&lt;/strong&gt;, then &lt;strong&gt;Event history&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Filter by &lt;strong&gt;Event name&lt;/strong&gt; and search for &lt;code&gt;CreateTopic&lt;/code&gt;. That is the SNS topic from my last post.&lt;/li&gt;
&lt;li&gt;Open the event and read the &lt;strong&gt;user name&lt;/strong&gt;, &lt;strong&gt;event time&lt;/strong&gt; and &lt;strong&gt;source IP address&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Create a trail: choose &lt;strong&gt;Create trail&lt;/strong&gt;, name it &lt;code&gt;management-events&lt;/code&gt;, let it create a new S3 bucket, and keep &lt;strong&gt;Management events&lt;/strong&gt; selected.&lt;/li&gt;
&lt;li&gt;Turn on &lt;strong&gt;Apply trail to all Regions&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Check it from the CLI: &lt;code&gt;aws cloudtrail lookup-events --lookup-attributes AttributeKey=EventName,AttributeValue=CreateTopic --max-results 5&lt;/code&gt;
&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Treating Event history as permanent when it only keeps 90 days.&lt;/li&gt;
&lt;li&gt;Creating a trail for one Region only.&lt;/li&gt;
&lt;li&gt;Turning on data events everywhere without thinking about cost.&lt;/li&gt;
&lt;li&gt;Leaving the log bucket open to more people than needed.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;CloudTrail answers who did what, when and from where.&lt;/li&gt;
&lt;li&gt;Event history is free and quick, a trail keeps the history.&lt;/li&gt;
&lt;li&gt;Use a trail that covers all Regions.&lt;/li&gt;
&lt;li&gt;Protect the bucket that holds the logs.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at AWS Config, which tracks how your resources looked over time.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>I built a GitOps pipeline with Jenkins, ArgoCD and Terraform on EKS: how it works and what broke</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Mon, 16 Nov 2026 04:27:07 +0000</pubDate>
      <link>https://dev.to/robinsishodia/i-built-a-gitops-pipeline-with-jenkins-argocd-and-terraform-on-eks-how-it-works-and-what-broke-5gof</link>
      <guid>https://dev.to/robinsishodia/i-built-a-gitops-pipeline-with-jenkins-argocd-and-terraform-on-eks-how-it-works-and-what-broke-5gof</guid>
      <description>&lt;p&gt;I wanted one project where a &lt;code&gt;git push&lt;/code&gt; ends up as a running pod with no manual steps, then the same thing on real AWS. I built it as a beginner, step by step, and wrote down what actually broke. Code and screenshots: &lt;a href="https://github.com/RobinSishodia/aws-eks-gitops-argocd" rel="noopener noreferrer"&gt;https://github.com/RobinSishodia/aws-eks-gitops-argocd&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What it does
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;I push a change to GitHub.&lt;/li&gt;
&lt;li&gt;Jenkins polls the repo (every 2 minutes), builds a Docker image and pushes it to Docker Hub.&lt;/li&gt;
&lt;li&gt;Jenkins commits the new image tag into &lt;code&gt;k8s/deployment.yaml&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;ArgoCD sees the manifest change and syncs it to Kubernetes (automated sync, prune and self-heal).&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;I proved it end to end by changing the heading in the Flask app and pushing. A few minutes later the new heading was served by freshly rolled-out pods, with no click after the push.&lt;/p&gt;

&lt;h2&gt;
  
  
  Repo layout
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;app/            Flask app + Dockerfile
k8s/            deployment.yaml, service.yaml
argocd-app.yaml ArgoCD Application (path: k8s, automated sync)
Jenkinsfile     build, push, update manifest
terraform/      VPC + EKS (modules)
docs/           screenshots
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  The Jenkins pipeline
&lt;/h2&gt;

&lt;p&gt;The pipeline has three stages: build the image, push it, then update the manifest and push that commit back to GitHub. The tag is the Jenkins build number.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight groovy"&gt;&lt;code&gt;&lt;span class="n"&gt;pipeline&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="n"&gt;agent&lt;/span&gt; &lt;span class="n"&gt;any&lt;/span&gt;
  &lt;span class="n"&gt;environment&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="no"&gt;IMAGE&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="s1"&gt;'robinsishodia/gitops-demo'&lt;/span&gt;
    &lt;span class="no"&gt;TAG&lt;/span&gt;   &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="s2"&gt;"${env.BUILD_NUMBER}"&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="n"&gt;stages&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="nf"&gt;stage&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'CI'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
      &lt;span class="k"&gt;when&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
        &lt;span class="ow"&gt;not&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="n"&gt;expression&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="nf"&gt;sh&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;script&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s1"&gt;'git log -1 --pretty=%B'&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;returnStdout&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kc"&gt;true&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;contains&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'[skip ci]'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
      &lt;span class="p"&gt;}&lt;/span&gt;
      &lt;span class="n"&gt;stages&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
        &lt;span class="nf"&gt;stage&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'Build image'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="n"&gt;steps&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="n"&gt;sh&lt;/span&gt; &lt;span class="s1"&gt;'docker build -t $IMAGE:$TAG app/'&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
        &lt;span class="nf"&gt;stage&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'Push image'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="cm"&gt;/* docker login with the 'dockerhub' credential, then docker push */&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
        &lt;span class="nf"&gt;stage&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'Update manifest'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
          &lt;span class="n"&gt;steps&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
            &lt;span class="c1"&gt;// GitHub credential injected with withCredentials&lt;/span&gt;
            &lt;span class="n"&gt;sh&lt;/span&gt; &lt;span class="s2"&gt;'''
              sed -i "s#image: .*#image: $IMAGE:$TAG#" k8s/deployment.yaml
              git add k8s/deployment.yaml
              git commit -m "ci: deploy $IMAGE:$TAG [skip ci]" || exit 0
              git push origin HEAD:main
            '''&lt;/span&gt;
          &lt;span class="p"&gt;}&lt;/span&gt;
        &lt;span class="p"&gt;}&lt;/span&gt;
      &lt;span class="p"&gt;}&lt;/span&gt;
    &lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The &lt;code&gt;when&lt;/code&gt; block is the loop guard, explained below. Credentials live in Jenkins, never in the repo.&lt;/p&gt;

&lt;h2&gt;
  
  
  The ArgoCD side
&lt;/h2&gt;

&lt;p&gt;ArgoCD watches the &lt;code&gt;k8s/&lt;/code&gt; folder and keeps the cluster equal to Git. ArgoCD polls Git about every 3 minutes. If you do not want to wait, force a refresh:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;kubectl -n argocd annotate application gitops-demo argocd.argoproj.io/refresh=hard --overwrite
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;That is how I saw the old pods (12 minutes old) replaced by new ones within about a minute.&lt;/p&gt;

&lt;h2&gt;
  
  
  Terraform and EKS
&lt;/h2&gt;

&lt;p&gt;I rebuilt the same cluster on AWS with Terraform: a VPC (2 AZs, public and private subnets, a single NAT gateway) and an EKS cluster with a managed node group of 2 nodes. I installed ArgoCD on it, applied the same Application manifest, and watched it come up Synced and Healthy. Then I ran &lt;code&gt;terraform destroy&lt;/code&gt; (56 resources), checked in the console and CLI that nothing was left, and deleted the IAM access key, so the lab cost nothing beyond free-plan credits.&lt;/p&gt;

&lt;p&gt;The workflow I used for safety: &lt;code&gt;terraform plan -out=tfplan&lt;/code&gt;, read the plan, then &lt;code&gt;terraform apply tfplan&lt;/code&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What went wrong, and the fixes
&lt;/h2&gt;

&lt;h3&gt;
  
  
  1. Jenkins rebuilt its own commits
&lt;/h3&gt;

&lt;p&gt;The pipeline pushes a commit, so polling would trigger another build, forever. Jenkins core does not honour &lt;code&gt;[skip ci]&lt;/code&gt;, so I wrapped the stages in a &lt;code&gt;when&lt;/code&gt; guard that skips a build whose latest commit message contains it (see the Jenkinsfile above). I also tested it: one build ran fully, the next was skipped.&lt;/p&gt;

&lt;p&gt;A self-inflicted detail: I once wrote the literal text &lt;code&gt;[skip ci]&lt;/code&gt; inside my own commit message while explaining the guard, and Jenkins correctly skipped a build I wanted. I fixed it with an empty commit with a neutral message.&lt;/p&gt;

&lt;h3&gt;
  
  
  2. AWS Free plan instance types
&lt;/h3&gt;

&lt;p&gt;The node group failed with &lt;code&gt;t3.medium&lt;/code&gt; not being free-tier eligible. On the Free plan, only certain instance types are allowed. I listed them with the EC2 CLI filtered by free-tier eligibility and switched to &lt;code&gt;c7i-flex.large&lt;/code&gt;.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. Nodes that never joined the cluster
&lt;/h3&gt;

&lt;p&gt;This one cost the most time. The node group sat on "Still creating" for hours. The cause: in version 21 of the terraform-aws-modules EKS module, the self-managed default add-ons are not bootstrapped, so vpc-cni, kube-proxy and CoreDNS were never installed and the nodes had no networking. The fix was to declare them:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight hcl"&gt;&lt;code&gt;&lt;span class="n"&gt;addons&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="n"&gt;coredns&lt;/span&gt;                &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="n"&gt;eks&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;pod&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;identity&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;agent&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="n"&gt;before_compute&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="kc"&gt;true&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="n"&gt;kube&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;proxy&lt;/span&gt;             &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="n"&gt;vpc&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;cni&lt;/span&gt;                &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="n"&gt;before_compute&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="kc"&gt;true&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;I also added &lt;code&gt;timeouts = { create = "25m" }&lt;/code&gt; on the node group so a stuck build fails after 25 minutes instead of running overnight.&lt;/p&gt;

&lt;h3&gt;
  
  
  4. Small things that cost me time
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;Port 8082 was already used by Jenkins, so I used 8090 for the app and 9443 for ArgoCD port-forwards.&lt;/li&gt;
&lt;li&gt;My Terraform plan files were not covered by &lt;code&gt;*.tfplan&lt;/code&gt; in &lt;code&gt;.gitignore&lt;/code&gt;; I added &lt;code&gt;tfplan*&lt;/code&gt; before committing.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;kubectl&lt;/code&gt; had two contexts (EKS and kind). Always check &lt;code&gt;kubectl config current-context&lt;/code&gt; before running commands.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  What I learned
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;GitOps is mostly discipline: the cluster changes only through Git.&lt;/li&gt;
&lt;li&gt;Any pipeline that commits to its own repo needs a loop guard.&lt;/li&gt;
&lt;li&gt;Read the module docs for what is not installed by default.&lt;/li&gt;
&lt;li&gt;Always destroy lab infrastructure, verify it, and delete the access keys.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  What is next
&lt;/h2&gt;

&lt;p&gt;A real GitHub webhook instead of polling (Jenkins runs on my laptop, so it needs a public tunnel or a hosted Jenkins), then monitoring with Prometheus and Grafana.&lt;/p&gt;

&lt;p&gt;If you spot something I could do better, I would like to hear it in the comments.&lt;/p&gt;

</description>
      <category>devops</category>
      <category>kubernetes</category>
      <category>terraform</category>
      <category>aws</category>
    </item>
    <item>
      <title>AWS CloudWatch Alarms and SNS: Get Notified When Something Goes Wrong</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Sun, 15 Nov 2026 16:36:42 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-cloudwatch-alarms-and-sns-get-notified-when-something-goes-wrong-47d</link>
      <guid>https://dev.to/robinsishodia/aws-cloudwatch-alarms-and-sns-get-notified-when-something-goes-wrong-47d</guid>
      <description>&lt;p&gt;In the last two posts I set up several accounts and a way for people to sign in. Now the question every real environment hits: when something goes wrong, &lt;strong&gt;who finds out, and how fast?&lt;/strong&gt; The basic answer on AWS is a &lt;strong&gt;CloudWatch alarm&lt;/strong&gt; that sends a message through &lt;strong&gt;Amazon SNS&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What each service does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Amazon CloudWatch&lt;/strong&gt; collects metrics, such as CPU use or estimated charges, and can watch them with alarms.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;CloudWatch alarm&lt;/strong&gt; compares a metric to a threshold and moves between three states: &lt;strong&gt;OK&lt;/strong&gt;, &lt;strong&gt;ALARM&lt;/strong&gt; and &lt;strong&gt;INSUFFICIENT_DATA&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Amazon SNS&lt;/strong&gt; (Simple Notification Service) sends a message to everyone subscribed to a &lt;strong&gt;topic&lt;/strong&gt;, by email, SMS or other targets.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The building blocks
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Term&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Metric&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A number measured over time&lt;/td&gt;
&lt;td&gt;Example: CPUUtilization or EstimatedCharges&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Alarm&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A rule on one metric&lt;/td&gt;
&lt;td&gt;Threshold, period and number of periods&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;SNS topic&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A channel for messages&lt;/td&gt;
&lt;td&gt;Alarms publish to it&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Subscription&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Who receives the topic&lt;/td&gt;
&lt;td&gt;An email subscription must be confirmed&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Alarm action&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;What the alarm does on a state change&lt;/td&gt;
&lt;td&gt;Often: publish to an SNS topic&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Why a topic and not a direct email
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;One topic can notify a whole team, and you can add people without touching the alarm.&lt;/li&gt;
&lt;li&gt;The same topic can be reused by many alarms.&lt;/li&gt;
&lt;li&gt;Later you can add other targets, such as a chat tool or a Lambda function.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: a billing alert in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Sign in with a user that can open the Billing console, and turn on &lt;strong&gt;Receive CloudWatch billing alerts&lt;/strong&gt; in Billing preferences.&lt;/li&gt;
&lt;li&gt;Switch the Region to &lt;strong&gt;US East (N. Virginia)&lt;/strong&gt;, because billing metrics are only published there.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;SNS&lt;/strong&gt;, create a &lt;strong&gt;Standard&lt;/strong&gt; topic named &lt;code&gt;billing-alerts&lt;/code&gt;, and add an &lt;strong&gt;Email&lt;/strong&gt; subscription with your address.&lt;/li&gt;
&lt;li&gt;Open the email and choose &lt;strong&gt;Confirm subscription&lt;/strong&gt;. Until you do, nothing is delivered.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;CloudWatch&lt;/strong&gt;, then &lt;strong&gt;Alarms&lt;/strong&gt;, then &lt;strong&gt;Create alarm&lt;/strong&gt;. Select the metric &lt;strong&gt;Billing &amp;gt; Total Estimated Charge&lt;/strong&gt; and set a threshold such as 5 USD.&lt;/li&gt;
&lt;li&gt;For the alarm action choose &lt;strong&gt;In alarm&lt;/strong&gt;, send to the &lt;code&gt;billing-alerts&lt;/code&gt; topic, and create the alarm.&lt;/li&gt;
&lt;li&gt;Test the notification path with the CLI: &lt;code&gt;aws cloudwatch set-alarm-state --alarm-name &amp;lt;your-alarm&amp;gt; --state-value ALARM --state-reason "test"&lt;/code&gt; and check that the email arrives.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Forgetting to confirm the email subscription.&lt;/li&gt;
&lt;li&gt;Creating the billing alarm in a Region other than US East (N. Virginia).&lt;/li&gt;
&lt;li&gt;Setting the threshold so high that the alarm never fires.&lt;/li&gt;
&lt;li&gt;Sending alerts to one personal inbox that nobody else watches.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;An alarm watches a metric, SNS delivers the message.&lt;/li&gt;
&lt;li&gt;Use a topic so you can add people and targets later.&lt;/li&gt;
&lt;li&gt;Confirm subscriptions and test the path on purpose.&lt;/li&gt;
&lt;li&gt;Start with a billing alarm in every new account.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at how to see who did what in your account, with AWS CloudTrail.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS IAM Identity Center: Single Sign-On Across Your AWS Accounts</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Sat, 14 Nov 2026 12:14:05 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-iam-identity-center-single-sign-on-across-your-aws-accounts-2j1k</link>
      <guid>https://dev.to/robinsishodia/aws-iam-identity-center-single-sign-on-across-your-aws-accounts-2j1k</guid>
      <description>&lt;p&gt;Yesterday I set up an organization with several accounts. Now the next question: how do &lt;strong&gt;people&lt;/strong&gt; get into those accounts without a pile of IAM users and long-term access keys? The answer is &lt;strong&gt;AWS IAM Identity Center&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What it does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;It is the AWS service for &lt;strong&gt;workforce single sign-on&lt;/strong&gt; across the accounts in your organization and supported applications.&lt;/li&gt;
&lt;li&gt;People sign in once at the &lt;strong&gt;AWS access portal&lt;/strong&gt; and pick the account and role they need.&lt;/li&gt;
&lt;li&gt;Access is given through &lt;strong&gt;temporary credentials&lt;/strong&gt;, so there are no long-term keys to lose.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The building blocks
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Term&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Identity source&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Where users and groups live&lt;/td&gt;
&lt;td&gt;Built-in directory, Active Directory or an external identity provider&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Permission set&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A template of permissions&lt;/td&gt;
&lt;td&gt;Identity Center creates a role from it in each assigned account&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Assignment&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Links a user or group to an account with a permission set&lt;/td&gt;
&lt;td&gt;Assign groups, not single users&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Access portal&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;The sign-in page for your users&lt;/td&gt;
&lt;td&gt;One URL for all accounts&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  How it fits with Organizations
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Enable it in the &lt;strong&gt;management account&lt;/strong&gt; with AWS Organizations turned on.&lt;/li&gt;
&lt;li&gt;Choose the &lt;strong&gt;home Region&lt;/strong&gt; carefully, because it is hard to change later.&lt;/li&gt;
&lt;li&gt;Roles created from permission sets show up in each account with names starting &lt;code&gt;AWSReservedSSO_&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Session length is set per permission set.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;IAM Identity Center&lt;/strong&gt; in the management account and choose &lt;strong&gt;Enable&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Under &lt;strong&gt;Groups&lt;/strong&gt;, create a group named &lt;code&gt;ReadOnly&lt;/code&gt;, then create a user and add them to it.&lt;/li&gt;
&lt;li&gt;Under &lt;strong&gt;Permission sets&lt;/strong&gt;, create one that uses the AWS managed &lt;strong&gt;ViewOnlyAccess&lt;/strong&gt; or &lt;strong&gt;ReadOnlyAccess&lt;/strong&gt; policy.&lt;/li&gt;
&lt;li&gt;Under &lt;strong&gt;AWS accounts&lt;/strong&gt;, select an account, choose &lt;strong&gt;Assign users or groups&lt;/strong&gt;, and pick the group and the permission set.&lt;/li&gt;
&lt;li&gt;Open the &lt;strong&gt;access portal URL&lt;/strong&gt;, sign in as the user, and choose the account and role.&lt;/li&gt;
&lt;li&gt;Optional: run &lt;code&gt;aws configure sso&lt;/code&gt;, then &lt;code&gt;aws sso login&lt;/code&gt;, and check with &lt;code&gt;aws sts get-caller-identity&lt;/code&gt;.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Creating IAM users for every person instead of using Identity Center.&lt;/li&gt;
&lt;li&gt;Assigning permission sets to single users, which gets hard to manage.&lt;/li&gt;
&lt;li&gt;Giving broad access when a read-only permission set is enough.&lt;/li&gt;
&lt;li&gt;Forgetting MFA for the people who sign in.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Use Identity Center for people and IAM roles for workloads.&lt;/li&gt;
&lt;li&gt;Define access once in a permission set and assign it to groups.&lt;/li&gt;
&lt;li&gt;Prefer temporary credentials over long-term keys.&lt;/li&gt;
&lt;li&gt;Keep MFA on for every user.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at what happens when something goes wrong in your accounts, starting with how to be notified.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS Organizations: Multiple Accounts, OUs and Service Control Policies</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Fri, 13 Nov 2026 14:39:47 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-organizations-multiple-accounts-ous-and-service-control-policies-4f1m</link>
      <guid>https://dev.to/robinsishodia/aws-organizations-multiple-accounts-ous-and-service-control-policies-4f1m</guid>
      <description>&lt;p&gt;So far I have secured root, set a budget, turned on CloudTrail and CloudWatch, and added tags. All of that sits in one account. The next step is the structure around it: &lt;strong&gt;AWS Organizations&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why more than one account
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;An AWS account is a strong boundary for &lt;strong&gt;security, cost and blast radius&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;A mistake in a sandbox account cannot touch production.&lt;/li&gt;
&lt;li&gt;Each account has its own root user, IAM and limits, so teams can work separately.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The building blocks
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Term&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Organization&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;The group of accounts you manage together&lt;/td&gt;
&lt;td&gt;Created from one account&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Management account&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;The account that creates the organization and pays the bill&lt;/td&gt;
&lt;td&gt;SCPs do not apply to it&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Member accounts&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Every other account in the organization&lt;/td&gt;
&lt;td&gt;One per workload or team&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;OU&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;An organizational unit that groups accounts&lt;/td&gt;
&lt;td&gt;Policies on an OU apply to its accounts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;SCP&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;A service control policy&lt;/td&gt;
&lt;td&gt;Sets the maximum permissions, never grants them&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Consolidated billing
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;All member accounts roll up to &lt;strong&gt;one bill&lt;/strong&gt; on the management account.&lt;/li&gt;
&lt;li&gt;Usage across accounts can count together for &lt;strong&gt;volume discounts&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Tags still let you split cost by project or owner.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  How SCPs behave
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;An SCP is a &lt;strong&gt;guardrail&lt;/strong&gt;: an action must be allowed by the SCP &lt;strong&gt;and&lt;/strong&gt; by the identity's own IAM policy.&lt;/li&gt;
&lt;li&gt;They apply to IAM users and roles in member accounts, including the account root user.&lt;/li&gt;
&lt;li&gt;They do not apply to the &lt;strong&gt;management account&lt;/strong&gt;, so keep workloads out of it.&lt;/li&gt;
&lt;li&gt;Use &lt;strong&gt;all features&lt;/strong&gt; mode to turn SCPs on.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  A simple starting structure
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Security&lt;/strong&gt; OU: logging and audit accounts&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Workloads&lt;/strong&gt; OU: production and non-production accounts&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Sandbox&lt;/strong&gt; OU: learning and experiments, with tighter SCPs&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;AWS Organizations&lt;/strong&gt; and choose &lt;strong&gt;Create an organization&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Check that &lt;strong&gt;MFA is on for the management account root user&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Create an OU named &lt;strong&gt;Sandbox&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Choose &lt;strong&gt;Add an AWS account&lt;/strong&gt;, then &lt;strong&gt;Create an AWS account&lt;/strong&gt;, and give it a unique email address and a name. Move it into the Sandbox OU.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Policies&lt;/strong&gt;, enable service control policies, and attach a deny policy to the Sandbox OU, for example one that stops the account leaving the organization.&lt;/li&gt;
&lt;li&gt;Sign in to the management account and &lt;strong&gt;assume the OrganizationAccountAccessRole&lt;/strong&gt; in the new account to check it works.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Running workloads in the management account.&lt;/li&gt;
&lt;li&gt;Using SCPs as if they grant permissions. They only limit.&lt;/li&gt;
&lt;li&gt;Reusing one email address for several accounts. Each account needs a unique one.&lt;/li&gt;
&lt;li&gt;Attaching SCPs to the organization root before testing on a Sandbox OU.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Use separate accounts as security and cost boundaries.&lt;/li&gt;
&lt;li&gt;Group accounts into OUs and attach policy once.&lt;/li&gt;
&lt;li&gt;SCPs set the maximum, IAM still grants.&lt;/li&gt;
&lt;li&gt;Keep the management account for administration only.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at how people and tools sign in across these accounts.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
    <item>
      <title>AWS Tagging Strategy: Name, Owner and Environment Tags for Every Resource</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Tue, 10 Nov 2026 18:53:27 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-tagging-strategy-name-owner-and-environment-tags-for-every-resource-h8h</link>
      <guid>https://dev.to/robinsishodia/aws-tagging-strategy-name-owner-and-environment-tags-for-every-resource-h8h</guid>
      <description>&lt;p&gt;By now I have IAM, a budget, CloudTrail and CloudWatch. One more habit makes all of them easier to use: &lt;strong&gt;tags&lt;/strong&gt;. When the bill rises or something looks odd, tags tell you &lt;strong&gt;what the resource is, who owns it and what it is for&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What tags are
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;A tag is a &lt;strong&gt;key and value pair&lt;/strong&gt; attached to a resource, such as &lt;code&gt;Environment = dev&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Most AWS resources support tags, and you can have up to &lt;strong&gt;50 tags per resource&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Tags are labels for people and tools. They do not change how the resource works.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  A simple starting set
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Key&lt;/th&gt;
&lt;th&gt;Example value&lt;/th&gt;
&lt;th&gt;Why&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Name&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;&lt;code&gt;web-server-1&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Shows in the console list&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Environment&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;
&lt;code&gt;dev&lt;/code&gt;, &lt;code&gt;test&lt;/code&gt;, &lt;code&gt;prod&lt;/code&gt;
&lt;/td&gt;
&lt;td&gt;Separate workloads&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Owner&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;&lt;code&gt;robin&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Who to ask&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Project&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;&lt;code&gt;blog-site&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Group related resources&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;CostCenter&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;&lt;code&gt;learning&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Split the bill&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What tags are used for
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Cost allocation:&lt;/strong&gt; activate &lt;strong&gt;cost allocation tags&lt;/strong&gt; in Billing, then filter Cost Explorer by tag.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Access control:&lt;/strong&gt; IAM policies can allow or deny actions based on tags (ABAC).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Automation:&lt;/strong&gt; scripts can find resources by tag, such as all &lt;code&gt;Environment = dev&lt;/code&gt; instances to stop at night.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Search:&lt;/strong&gt; Tag Editor and Resource Groups find resources across Regions.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Good to know
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Tag keys and values are &lt;strong&gt;case sensitive&lt;/strong&gt;, so &lt;code&gt;Owner&lt;/code&gt; and &lt;code&gt;owner&lt;/code&gt; are different.&lt;/li&gt;
&lt;li&gt;Tags are &lt;strong&gt;not retroactive&lt;/strong&gt; in Cost Explorer. Cost allocation tags only show costs from the time you activate them.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Tag policies&lt;/strong&gt; in AWS Organizations help keep names consistent.&lt;/li&gt;
&lt;li&gt;Do not put secrets or personal data in tags.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;EC2&lt;/strong&gt; or &lt;strong&gt;S3&lt;/strong&gt; and select a resource you created.&lt;/li&gt;
&lt;li&gt;Add the tags &lt;strong&gt;Name&lt;/strong&gt;, &lt;strong&gt;Environment&lt;/strong&gt;, &lt;strong&gt;Owner&lt;/strong&gt; and &lt;strong&gt;Project&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Resource Groups &amp;amp; Tag Editor&lt;/strong&gt; and use &lt;strong&gt;Tag Editor&lt;/strong&gt; to search by &lt;code&gt;Environment = dev&lt;/code&gt; across Regions.&lt;/li&gt;
&lt;li&gt;Open &lt;strong&gt;Billing&lt;/strong&gt; and &lt;strong&gt;Cost allocation tags&lt;/strong&gt;, then activate &lt;code&gt;Project&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;After a day, filter &lt;strong&gt;Cost Explorer&lt;/strong&gt; by the &lt;code&gt;Project&lt;/code&gt; tag.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Inconsistent keys, such as &lt;code&gt;env&lt;/code&gt;, &lt;code&gt;Env&lt;/code&gt; and &lt;code&gt;Environment&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Tagging only some resources, so reports have gaps.&lt;/li&gt;
&lt;li&gt;Forgetting to activate cost allocation tags.&lt;/li&gt;
&lt;li&gt;Storing sensitive data in tag values.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Tags answer what, who and why for every resource.&lt;/li&gt;
&lt;li&gt;Pick a small standard set and use it everywhere.&lt;/li&gt;
&lt;li&gt;Activate cost allocation tags to see costs by tag.&lt;/li&gt;
&lt;li&gt;Keep tag names consistent.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at how to see all of this from a single place: AWS Organizations and account structure.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>tagging</category>
      <category>cloud</category>
      <category>devops</category>
    </item>
    <item>
      <title>AWS CloudWatch Basics: Metrics, Logs and Alarms for Your Resources</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Mon, 09 Nov 2026 16:15:52 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-cloudwatch-basics-metrics-logs-and-alarms-for-your-resources-3d0n</link>
      <guid>https://dev.to/robinsishodia/aws-cloudwatch-basics-metrics-logs-and-alarms-for-your-resources-3d0n</guid>
      <description>&lt;p&gt;Last time I looked at &lt;strong&gt;who&lt;/strong&gt; changed things with CloudTrail. The next question is different: &lt;strong&gt;what is happening inside my resources right now&lt;/strong&gt;? That is &lt;strong&gt;Amazon CloudWatch&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What CloudWatch does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Metrics:&lt;/strong&gt; numbers over time, such as EC2 CPU utilization or S3 request counts.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Logs:&lt;/strong&gt; log data from your applications and services, stored in log groups.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Alarms:&lt;/strong&gt; watch a metric and do something when it crosses a threshold.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Dashboards:&lt;/strong&gt; put the graphs you care about on one page.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  CloudTrail vs CloudWatch
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Question&lt;/th&gt;
&lt;th&gt;Service&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Who made this change?&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;CloudTrail&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Is this resource healthy and busy?&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;CloudWatch&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  How an alarm works
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;An alarm watches &lt;strong&gt;one metric&lt;/strong&gt; over a &lt;strong&gt;period&lt;/strong&gt; (for example, 5 minutes).&lt;/li&gt;
&lt;li&gt;Its state is &lt;strong&gt;OK&lt;/strong&gt;, &lt;strong&gt;ALARM&lt;/strong&gt; or &lt;strong&gt;INSUFFICIENT_DATA&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;When the state changes, it can send a message to an &lt;strong&gt;SNS topic&lt;/strong&gt;, which can email you.&lt;/li&gt;
&lt;li&gt;The &lt;strong&gt;threshold&lt;/strong&gt; and the number of &lt;strong&gt;periods&lt;/strong&gt; to evaluate decide how sensitive it is.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Good to know
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;EC2 sends basic metrics every &lt;strong&gt;5 minutes&lt;/strong&gt; by default, and &lt;strong&gt;detailed monitoring&lt;/strong&gt; gives 1-minute data (extra cost).&lt;/li&gt;
&lt;li&gt;EC2 does &lt;strong&gt;not&lt;/strong&gt; report memory or disk usage by default. You need the &lt;strong&gt;CloudWatch agent&lt;/strong&gt; for those.&lt;/li&gt;
&lt;li&gt;Logs stay forever unless you set a &lt;strong&gt;retention&lt;/strong&gt; period, so set one.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;CloudWatch&lt;/strong&gt; and browse &lt;strong&gt;All metrics&lt;/strong&gt; for EC2 or S3.&lt;/li&gt;
&lt;li&gt;Choose &lt;strong&gt;Alarms&lt;/strong&gt;, then &lt;strong&gt;Create alarm&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Pick a metric, such as &lt;strong&gt;EC2 CPUUtilization&lt;/strong&gt; for one instance.&lt;/li&gt;
&lt;li&gt;Set a threshold, for example greater than 70% for 2 periods.&lt;/li&gt;
&lt;li&gt;Create or choose an &lt;strong&gt;SNS topic&lt;/strong&gt; with your email, and confirm the subscription email.&lt;/li&gt;
&lt;li&gt;Save the alarm and check that it shows &lt;strong&gt;OK&lt;/strong&gt; or &lt;strong&gt;INSUFFICIENT_DATA&lt;/strong&gt; at first.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Creating alarms with no action, so nobody is told.&lt;/li&gt;
&lt;li&gt;Forgetting to confirm the SNS email subscription.&lt;/li&gt;
&lt;li&gt;Expecting memory metrics without the CloudWatch agent.&lt;/li&gt;
&lt;li&gt;Leaving log groups on never-expire retention.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;CloudTrail tells you who changed it, CloudWatch tells you how it is doing.&lt;/li&gt;
&lt;li&gt;Metrics feed alarms, alarms feed SNS, SNS reaches you.&lt;/li&gt;
&lt;li&gt;Install the agent for memory and disk.&lt;/li&gt;
&lt;li&gt;Set log retention.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at how to use tags to keep resources organized and costs traceable.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloudwatch</category>
      <category>monitoring</category>
      <category>devops</category>
    </item>
    <item>
      <title>AWS CloudTrail Basics: Who Did What, When and From Where</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Sun, 08 Nov 2026 17:56:22 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-cloudtrail-basics-who-did-what-when-and-from-where-22ad</link>
      <guid>https://dev.to/robinsishodia/aws-cloudtrail-basics-who-did-what-when-and-from-where-22ad</guid>
      <description>&lt;p&gt;This week I locked the root account and set a budget. Now the question that comes after both: when something changes in the account, how do I find out &lt;strong&gt;who did it, what they did and when&lt;/strong&gt;? That is &lt;strong&gt;AWS CloudTrail&lt;/strong&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What CloudTrail records
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;CloudTrail logs &lt;strong&gt;API activity&lt;/strong&gt; in your account, whether it came from the console, the CLI, an SDK or another AWS service.&lt;/li&gt;
&lt;li&gt;Each event says &lt;strong&gt;who&lt;/strong&gt; (the identity), &lt;strong&gt;what&lt;/strong&gt; (the API call), &lt;strong&gt;when&lt;/strong&gt;, and &lt;strong&gt;where from&lt;/strong&gt; (the source IP address).&lt;/li&gt;
&lt;li&gt;It is the audit trail you open when a resource disappears or a setting changes.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Event types
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Type&lt;/th&gt;
&lt;th&gt;What it covers&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Management events&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Control-plane actions, such as creating a role or launching an instance&lt;/td&gt;
&lt;td&gt;Recorded by default&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Data events&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Object-level actions, such as S3 object reads or Lambda invokes&lt;/td&gt;
&lt;td&gt;Off by default, extra cost&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Insights events&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Unusual API activity&lt;/td&gt;
&lt;td&gt;Optional&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Event history vs a trail
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Event history&lt;/strong&gt; is on by default and lets you look up recent management events (the last 90 days) in the console, with no setup.&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;trail&lt;/strong&gt; delivers events to an &lt;strong&gt;S3 bucket&lt;/strong&gt; (and optionally CloudWatch Logs) so you keep them longer and can search them with other tools.&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;multi-Region trail&lt;/strong&gt; records activity in all Regions, so nothing slips through in a Region you do not use.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Open &lt;strong&gt;CloudTrail&lt;/strong&gt; and choose &lt;strong&gt;Event history&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Filter by &lt;strong&gt;Event name&lt;/strong&gt; &lt;code&gt;ConsoleLogin&lt;/code&gt; and open one event to see who signed in and from which IP address.&lt;/li&gt;
&lt;li&gt;Filter by &lt;strong&gt;User name&lt;/strong&gt; &lt;code&gt;root&lt;/code&gt; to check whether root has been used, tying back to the root account checklist.&lt;/li&gt;
&lt;li&gt;Choose &lt;strong&gt;Create trail&lt;/strong&gt;, name it, and create or choose an &lt;strong&gt;S3 bucket&lt;/strong&gt; for the logs.&lt;/li&gt;
&lt;li&gt;Enable a &lt;strong&gt;multi-Region trail&lt;/strong&gt; and &lt;strong&gt;log file validation&lt;/strong&gt;, then create it.&lt;/li&gt;
&lt;li&gt;After a few minutes, open the S3 bucket and find the delivered log files.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Relying on event history only, then needing events older than 90 days.&lt;/li&gt;
&lt;li&gt;Creating a trail in one Region and missing activity elsewhere.&lt;/li&gt;
&lt;li&gt;Leaving the log bucket open. Keep it private.&lt;/li&gt;
&lt;li&gt;Turning on data events everywhere without checking the cost.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;CloudTrail answers who, what, when and from where.&lt;/li&gt;
&lt;li&gt;Event history is free to browse, a trail keeps the records.&lt;/li&gt;
&lt;li&gt;Use a multi-Region trail and keep the bucket private.&lt;/li&gt;
&lt;li&gt;Check root activity regularly.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll look at how to see what is happening inside your resources, not just who changed them.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloudtrail</category>
      <category>security</category>
      <category>devops</category>
    </item>
    <item>
      <title>AWS Budgets and Billing Alerts: Set a Spending Limit Before You Build</title>
      <dc:creator>Robin Sishodia</dc:creator>
      <pubDate>Sat, 07 Nov 2026 17:25:42 +0000</pubDate>
      <link>https://dev.to/robinsishodia/aws-budgets-and-billing-alerts-set-a-spending-limit-before-you-build-pfg</link>
      <guid>https://dev.to/robinsishodia/aws-budgets-and-billing-alerts-set-a-spending-limit-before-you-build-pfg</guid>
      <description>&lt;p&gt;Yesterday I locked down the root account and noted "keep billing alerts up to date". Today is the other half of the front door: making sure a forgotten resource never turns into a surprise bill. &lt;strong&gt;AWS Budgets&lt;/strong&gt; does that.&lt;/p&gt;

&lt;h2&gt;
  
  
  What AWS Budgets does
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;You set a &lt;strong&gt;spending limit&lt;/strong&gt; (or a usage limit) and AWS compares your actual and forecasted spend against it.&lt;/li&gt;
&lt;li&gt;When a threshold is crossed, it &lt;strong&gt;emails you&lt;/strong&gt; (and can notify an SNS topic).&lt;/li&gt;
&lt;li&gt;Spend data is not instant. Billing data can lag, so treat alerts as an early warning, not a live meter.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Types of budget
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Type&lt;/th&gt;
&lt;th&gt;Tracks&lt;/th&gt;
&lt;th&gt;Good for&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Cost budget&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Dollars spent&lt;/td&gt;
&lt;td&gt;Almost every learner&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Usage budget&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Units used, such as hours or GB&lt;/td&gt;
&lt;td&gt;Watching a specific service&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Savings Plans / Reservation budgets&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Coverage and utilization&lt;/td&gt;
&lt;td&gt;Later, when you commit to spend&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Two templates to start with
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Zero spend budget:&lt;/strong&gt; alerts you as soon as your spend goes above the free tier. Ideal for a learning account.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Monthly cost budget:&lt;/strong&gt; you choose a monthly amount, and the console sets alerts at &lt;strong&gt;85% of actual spend&lt;/strong&gt;, &lt;strong&gt;100% of actual spend&lt;/strong&gt; and &lt;strong&gt;100% of forecasted spend&lt;/strong&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Hands-on: try it in your own account
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Sign in and open &lt;strong&gt;Billing and Cost Management&lt;/strong&gt;, then &lt;strong&gt;Budgets&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Choose &lt;strong&gt;Create budget&lt;/strong&gt; and pick a &lt;strong&gt;template&lt;/strong&gt;, such as &lt;strong&gt;Zero spend budget&lt;/strong&gt; or &lt;strong&gt;Monthly cost budget&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Name it, set the amount (for example 10 USD) and add the &lt;strong&gt;email addresses&lt;/strong&gt; to alert.&lt;/li&gt;
&lt;li&gt;Review the thresholds and create the budget.&lt;/li&gt;
&lt;li&gt;Optionally, as the root user, turn on &lt;strong&gt;IAM user and role access to Billing information&lt;/strong&gt; so your admin identity can see costs without using root.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Common mistakes
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Creating a budget but typing the email address wrongly.&lt;/li&gt;
&lt;li&gt;Setting only an actual-spend alert, so you hear about a problem after the money is gone.&lt;/li&gt;
&lt;li&gt;Forgetting that some services keep charging until you delete the resource.&lt;/li&gt;
&lt;li&gt;Using root every time to check the bill instead of enabling billing access for IAM.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Worth knowing next
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Cost Explorer&lt;/strong&gt; shows where your spend comes from, and &lt;strong&gt;Cost Anomaly Detection&lt;/strong&gt; can flag unusual spikes. Budgets tells you &lt;em&gt;that&lt;/em&gt; spend is rising. Those two help explain &lt;em&gt;why&lt;/em&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Create a budget before you create any real resources.&lt;/li&gt;
&lt;li&gt;Alert on &lt;strong&gt;forecasted&lt;/strong&gt; spend as well as actual spend.&lt;/li&gt;
&lt;li&gt;Give billing access to an IAM admin, not just root.&lt;/li&gt;
&lt;li&gt;Delete what you stop using.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Tomorrow I'll go one step further into securing access to the account.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://sishodiarobin.hashnode.dev/aws-budgets-and-billing-alerts-set-a-spending-limit-before-you-build" rel="noopener noreferrer"&gt;my blog&lt;/a&gt;. Part of #120DaysOfAWS.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>aws</category>
      <category>cloud</category>
      <category>devops</category>
      <category>beginners</category>
    </item>
  </channel>
</rss>
