Releases · cloudposse/atmos

@aknysh

Improve logging and error handling @aknysh (#1358)

what

Improve logging
- Remove the deprecated logs functions throughout the codebase
- Only use the structured log functions (log.Info, log.Debug, log.Warn)
Improve error handling
- Add a new Go package errors with all functions and data structures related to error handling in the codebase (all other packages import the errors package and call its functions)
- Remove all functions from the codebase that would exit the program execution earlier: panic(), log.Fatal(), os.Exit(), and some old Atmos error handling functions
- In the new errors package, add a function to print errors in Markdown format (if Markdown is configured and TTY is supported), and a function to print errors and exit the program execution
Don't process Atmos YAML functions and Go templates in atmos validate stacks

why

Improve logging
- Standardize logs
- Make it easy to parse and analyze by log aggregators, SIEM tools, or monitoring systems
- Easier integration with tracing/Telemetry
- Consistency across logs
- Better debugging and contextual info
- Improved querying and filtering
Improve error handling
- Better, standard UI/UX
- All errors and program exits are handled in just one place, allowing easily configuring the look and feel of all error messages, and adding hooks for Tracing/Telemetry with just a few lines of code/config
- All early exits from any functions in the codebase (due to errors or other conditions) are now funneled through just one function in the errors package, allowing easier controlling the code execution and exit behavior
Don't process Atmos YAML functions and Go templates in atmos validate stacks. YAML functions and Go templates can access backends and clouds, but atmos validate stacks should not require authentication and assuming roles

test

atmos terraform apply --affected --all --dry-run

atmos terraform apply component-1 -s nonprods

atmos terraform apply component-2 -s prod

atmos terraform apply component-1 -s nonprod

atmos terraform apply --all --logs-level=Debu

Add pager to `atmos describe dependents` command @samtholiya (#1295)

what

Add pager to atmos describe dependents command

why

Helps to view the dependents better in case of large dependents list

Update help with pager @samtholiya (#1276)

what

Updated help with pager

why

Makes easier for users to view the content

Add `--format` flag to `atmos version` command @samtholiya (#1322)

what

Use different output formats: plain text, JSON, or YAML

why

Required for automation

@aknysh

Terraform/OpenTofu Multi-Component Commands (Filtered/Bulk Operations) @aknysh (#1327)

what

Add Terraform/OpenTofu Multi-Component Commands (Filtered/Bulk Operations)
- atmos terraform plan/apply/deploy --all
- atmos terraform plan/apply/deploy --all --stack <stack>
- atmos terraform plan/apply/deploy --affected
- atmos terraform plan/apply/deploy --affected --stack <stack>
- atmos terraform plan/apply/deploy --affected --include-dependents
- atmos terraform plan/apply/deploy --affected --stack <stack> --include-dependents
- atmos terraform plan/apply/deploy --components <component1>,<component2>
- atmos terraform plan/apply/deploy --components <component1>,<component2> --stack <stack>
- atmos terraform plan/apply/deploy --query <yq-expression>
- atmos terraform plan/apply/deploy --query <yq-expression> --stack <stack>
Add unit tests
Update docs
- https://atmos.tools/cli/commands/terraform/usage/
- https://atmos.tools/cli/commands/terraform/usage/#multi-component-commands-bulk-operations-examples

why

Allow Terraform/Opentofu operations (plan, apply, deploy) on a set of components in all stacks or a specific stack depending on stack, set of components, query, or filters

description

Atmos Terraform/OpenTofu commands fall into two categories:

Single-Component: Run Terraform for one component at a time
Multi-Component (Filtered/Bulk): Run Terraform across multiple components using stack names, selectors, or change detection

Single-Component Commands Usage

# Execute `terraform <command>` on a `component` in a `stack`
atmos terraform <command> <component> -s <stack> [options]
atmos terraform <command> <component> --stack <stack> [options]

Multi-Component Commands (Bulk Operations) Usage

# Execute `terraform <command>` on all components in the stack `prod`
atmos terraform <command> --stack prod

# Execute `terraform <command>` on components `component-1` and `component-2` in all stacks
atmos terraform <command> --components component-1,component-2

# Execute `terraform <command>` on components `component-1` and `component-2` in the stack `prod`
atmos terraform <command> --stack prod --components component-1,component-2

# Execute `terraform <command>` on all components in all stacks
atmos terraform <command> --all

# Execute `terraform <command>` on all components in the stack `prod`
atmos terraform <command> --all --stack prod

# Execute `terraform <command>` on all the directly affected components in all stacks in dependency order
# (if component dependencies are configured)
atmos terraform <command> --affected

# Execute `terraform <command>` on all the directly affected components in the `prod` stack in dependency order
# (if component dependencies are configured)
atmos terraform <command> --affected --stack prod

# Execute `terraform <command>` on all the directly affected components in all stacks in dependency order.
# For each directly affected component, detect the dependent components and process them in dependency order, recursively.
# Dependents are components that are indirectly affected, meaning that nothing in the current branch modifies their code
# or configs, but they are configured as dependencies of the components that are modified
atmos terraform <command> --affected --include-dependents

# Execute `terraform <command>` on all the directly affected components in the `prod` stack in dependency order.
# For each directly affected component, detect the dependent components and process them in dependency order, recursively.
atmos terraform <command> --affected --include-dependents --stack prod

# Execute `terraform <command>` on all components that have `vars.tags.team == "data"`, in all stacks
atmos terraform <command> --query '.vars.tags.team == "data"'

# Execute `terraform <command>` on all components that have `vars.tags.team == "eks"`, in the stack `prod`
atmos terraform <command> --query '.vars.tags.team == "eks"' --stack prod

# Execute `terraform <command>` on all components that have `settings.context.account_id == 12345`, in all stacks
atmos terraform <command> --query '.settings.context.account_id == 12345'

Multi-Component Commands (Bulk Operations) Examples

Let's assume that we have the following Atmos stack manifests in the prod and nonprod stacks,
with dependencies between the components:

components:
  terraform:
    vpc:
      vars:
        tags:
          # Team `network` manages the `vpc` component
          team: network
    eks/cluster:
      vars:
        tags:
          # Team `eks` manages the `eks/cluster` component
          team: eks
      settings:
        depends_on:
          # `eks/cluster` depends on the `vpc` component
          1:
            component: vpc
    eks/external-dns:
      vars:
        tags:
          # Team `eks` manages the `eks/external-dns` component
          team: eks
      settings:
        depends_on:
          # `eks/external-dns` depends on the `eks/cluster` component
          1:
            component: eks/cluster
    eks/karpenter:
      vars:
        tags:
          # Team `eks` manages the `eks/karpenter` component
          team: eks
      settings:
        depends_on:
          # `eks/karpenter` depends on the `eks/cluster` component
          1:
            component: eks/cluster
    eks/karpenter-node-pool:
      vars:
        tags:
          # Team `eks` manages the `eks/karpenter-node-pool` component
          team: eks
      settings:
        # `eks/karpenter-node-pool` depends on the `eks/cluster` and `eks/karpenter` components
        depends_on:
          1:
            component: eks/cluster
          2:
            component: eks/karpenter
    eks/istio/base:
      vars:
        tags:
          # Team `istio` manages the `eks/istio/base` component
          team: istio
      settings:
        # `eks/istio/base` depends on the `eks/cluster` component
        depends_on:
          1:
            component: eks/cluster
    eks/istio/istiod:
      vars:
        tags:
          # Team `istio` manages the `eks/istio/istiod` component
          team: istio
      settings:
        # `eks/istio/istiod` depends on the `eks/cluster` and `eks/istio/base` components
        depends_on:
          1:
            component: eks/cluster
          2:
            component: eks/istio/base
    eks/istio/test-app:
      vars:
        tags:
          # Team `istio` manages the `eks/istio/test-app` component
          team: istio
      settings:
        # `eks/istio/test-app` depends on the `eks/cluster`, `eks/istio/istiod` and `eks/istio/base` components
        depends_on:
          1:
            component: eks/cluster
          2:
            component: eks/istio/istiod
          3:
            component: eks/istio/base

Let's run the following Multi-Component commands in dry-run mode and review the output to understand what each command executes:

# Execute the `terraform apply` command on all components in all stacks

> atmos terraform apply --all --dry-run

Executing command="atmos terraform apply vpc -s nonprod"
Executing command="atmos terraform apply eks/cluster -s nonprod"
Executing command="atmos terraform apply eks/external-dns -s nonprod"
Executing command="atmos terraform apply eks/istio/base -s nonprod"
Executing command="atmos terraform apply eks/istio/istiod -s nonprod"
Executing command="atmos terraform apply eks/istio/test-app -s nonprod"
Executing command="atmos terraform apply eks/karpenter -s nonprod"
Executing command="atmos terraform apply eks/karpenter-node-pool -s nonprod"

Executing command="atmos terraform apply vpc -s prod"
Executing command="atmos terraform apply eks/cluster -s prod"
Executing command="atmos terraform apply eks/external-dns -s prod"
Executing command="atmos terraform apply eks/istio/base -s prod"
Executing command="atmos terraform apply eks/istio/istiod -s prod"
Executing command="atmos terraform apply eks/istio/test-app -s prod"
Executing command="atmos terraform apply eks/karpenter -s prod"
Executing command="atmos terraform apply eks/karpenter-node-pool -s prod"

# Execute the `terraform apply` command on all components in the `prod` stack

> atmos terraform apply --all --stack prod --dry-run

Executing command="atmos terraform apply vpc -s prod"
Executing command="atmos terraform apply eks/cluster -s prod"
Executing command="atmos terraform apply eks/external-dns -s prod"
Executing command="atmos terraform apply eks/istio/base -s prod"
Executing command="atmos terraform apply eks/istio/istiod -s prod"
Executing command="atmos terraform apply eks/istio/test-app -s prod"
Executing command="atmos terraform apply eks/karpenter -s prod"
Executing command="atmos terraform apply eks/karpenter-node-pool -s prod"

# Execute the `terraform apply` command on all components in the `prod` stack

> atmos terraform apply --stack prod --dry-run

Executing command="atmos terraform apply vpc -s prod"
Executing command="atmos terraform apply eks/cluster -s prod"
Executing command="atmos terraform apply eks/external-dns -s prod"
Executing command="atmos terraform apply eks/istio/base -s prod"
Executing command="atmos terraform apply eks/istio/istiod -s prod"
Executing command="atmos terraform apply eks/istio/test-app -s prod"
Executing command="atmos terraform apply eks/karpenter -s prod"
Executing command="atmos terraform apply eks/karpenter-node-pool -s prod"

# Execute the `terraform apply` command on the `vpc` and `eks/cluster` components
# in all stacks.

> atmos terraform apply --components vp...