MCA CLI

CLI to help automating MCA work

Installation

1. Clone repository
2. Run npm link

Local development

• Watch and compile on change npm run start
• Compile typescript to javascript with npm run build
• Run built mca cli with ./dist/bin/mca.js

Folder organization

• src (Contains source files)
  • bin (Starting point for cli app)
  • cmd (Command line configs using yargs commandDir)
  • lib (Code for to commands)
• assets (Assets required for the command line)
• dist (Build folder, same as src but with js files)

Linting

• Lint code with npm run lint
• Fix linter errors with npm run lint:fix

Testing

• Run tests with npm run test

Unit tests should be in the same location as the code with added spec.ts extension. Larger integration tests should be separated to test folder.

Release

Run npm run release to make version bump, add tags and update CHANGELOG automatically.

---

What is mca-cli?

mca-cli and mca-monitoring are used together to setup monitoring for resources in an AWS environment.

The mca monitoring command searches for resources in an AWS environment. It generates a Node project inside your main project folder and creates a config.yml, which lists all the resources, along with some default alarms.

How does monitoring work in AWS?

It is a combination of CloudWatch metrics and CloudWatch alarms.

A metric is a statistic. For example: AWS Lambda has an Invocations metric, which counts the number of times a function is invoked.

An alarm observes a single metric and initiates actions when a specified condition is met. The action could be sending a notification to a SNS topic.

---

Setting up default monitoring

1. Generate the monitoring folder

In the root project folder run

npx mca monitoring init -p <aws profile> -r <aws region> -o monitoring

Optional flags

--service - a space seperated list of service names to include in the search for resources. By default all resources are included:

• lambda
• dynamodb
• ecs
• apigateway
• cloudfront
• rds
• eks
• loggroup
• appsync
• sqs

--include: A list of regex patterns of resource names (or ids) to include in the monitoring By default all resources are included. Resources are identified by:

• (lambda) function name
• (dynamodb) table name
• (ecs) cluster name
• (apigateway) api name
• (cloudfront) distribution id or alias
• (rds) db instance identifier
• (eks) cluster name
• (appsync) api name
• (sqs) queue name

--exclude: Same as above, but resources are excluded.

--help See all options

2. Install NPM packages

In the monitoring folder run npm install

3. Deploy

In the monitoring folder run npm run deploy

***it

Customizing monitoring

Read more about CloudWatch concepts

Custom configurations should be listed in config.yml, under custom > default.

Configurable properties

• enabled Boolean.
  Whether to create an alarm for this metric.
• autoResolve Boolean
  (optional, default: false), Should the alarm automatically enter “OK” state.
• alarm
  • critical
    • comparisonOperator String
      (optional, default: GREATER_THAN_OR_EQUAL_TO_THRESHOLD). Comparison to use to check if metric is breaching. (available values)
    • threshold Number
      (required). The value against which the specified statistic is compared.
    • evaluationPeriods Nubmer
      (required). The number of periods over which data is compared to the specified threshold.
    • evaluateLowSampleCountPercentile Percentile
      (optional). Used only for alarms that are based on percentiles. Specifies whether to evaluate the data and potentially change the alarm state if there are too few data points to be statistically significant.
    • treatMissingData String
      (optional, default: NOT_BREACHING). Sets how this alarm is to handle missing data points. (available values)
• metric
  • period
    (optional, default: 5 minutes) The period over which the specified statistic is applied. Can have one of the following sub properties:
    • milliseconds Number
    • seconds Number
    • minutes Number
    • hours Number
    • days Number
    • isoString ISO 8601
  • statisticString
    (required, one of: Minimum, Maximum, Average, Sum, SampleCount, pNN.NN). What function to use for aggregating.
  • unitString
    (optional, default: undefined). Unit used to filter the metric stream. Only useful when datums are being emitted to the same metric stream under different units.

Log groups specific properties

• [custom metric name]
  • filter
    • pattern String
      (required). Filter pattern syntax.
      When using quotes for exact matches (e.g. “[ERROR]"), put single + double quotes (e.g. '"[ERROR]"'), or mca-monitoring will end up with a regex (e.g. [ERROR]).

Config.yml example

cli:
  version: 1
  services: # mca-cli will search for resources from these services
    - lambda
    - dynamodb
    - apigateway
    - cloudfront
    - loggroup
  includes: [] # Regex patterns (resource names) to include in monitoring. See "Optional flags" in "Setting up default monitoring" section above.
  excludes: # Exclude these resources patterns from monitoring.
    - '*ee*'
    - '*rapsiapp*'
    - '*dev*'
    - '*marketprice*'
    - '*warmup*'
    - '*error-handler*'
  profile: nc-personal-user # The AWS profile used during searching for resources with mca-cli and when deploying with mca-monitoring.
custom:
  default:
    lambda: # Config type. See "Config types and metrics" bellow.
      Errors: # Metric name
        enabled: true # Whether to create an alarm for this metric.
        autoresolve: false # Should the alarm automatically enter “OK” state.
          alarm:
            critical:
              comparisonOperator: GREATER_THAN_OR_EQUAL_TO_THRESHOLD # Comparison to use to check if metric is breaching.
              threshold: 1 # The value against which the specified statistic is compared.
              evaluationPeriods: 1 # The number of periods over which data is compared to the specified threshold.
          metric:
            period: # The period over which the specified statistic is applied.
              minutes: 15
            statistic: Minimum # What function to use for aggregating.
    cloudfront: # Config type
      4XXErrorRate: # Metric name
        enabled: false # Monitoring for this metric is disabled.
    logGroup:
      RuntimeErrors: # Custom metric name
        enabled: true
        alarm:
          critical:
            threshold: 1
            evaluationPeriods: 1
        metric:
          period:
            minutes: 5
          unit: Count
          statistic: Sum
        filter:
          pattern: ERROR -400 -401 -403 -404 -Timeout -DeprecationWarning
  snsTopic: # This topic is created by default and is used by all alarms.
    critical:
      name: Topic for mca monitoring alarms
      id: avena-alerts-alarm
      endpoints:
        - >-
          https://events.pagerduty.com/integration/58287e69892c4406aa88db8619721142/enqueue
      emails: []
lambdas: # Lambdas to be monitored.
  myTestLambda: {}
distributions: # CloudFront distributions to be monitored.
  E2K3LH1G46OF18: {}
  E3ADB61RBHAPW9: {}
  E35IJ0HST9PMZQ: {}
logGroups:
  /aws/lambda/avenakauppa-fi-analysis-prod-get-analysis: {}
  /aws/lambda/avenakauppa-fi-analysis-prod-post-analysis: {}

Config types and metrics

lambda

• Invocations
• Errors
• DeadLetterErrors
• DestinationDeliveryFailures
• Throttles
• ProvisionedConcurrencyInvocations
• ProvisionedConcurrencySpilloverInvocations
• Duration
• IteratorAge
• ConcurrencyExecutions
• ProvisionedConcurrencyExecutions
• ProvisionedConcurrencyUtilizations
• UnreservedConcurrentExecutions

table

• ConditionalCheckFailedRequests
• ConsumedReadCapacityUnits
• ConsumedWriteCapacityUnits
• MaxProvisionedTableReadCapacityUtilization
• MaxProvisionedTableWriteCapacityUtilization
• OnlineIndexConsumedWriteCapacity
• OnlineIndexPercentageProgress
• OnlineIndexThrottleEvents
• PendingReplicationCount
• ProvisionedReadCapacity
• ProvisionedWriteCapacity
• ReadThrottleEvents
• ReplicationLatency
• ReturnedBytes
• ReturnedItemCount
• ReturnedRecordsCount
• SystemErrors
• TimeToLiveDeletedItemCount
• ThrottledRequests
• TransactionConflict
• WriteThrottleEvents

account

This is part of the AWS/DynamoDB namespace

• UserErrors

clusters

• CPUReservation
• CPUUtilization
• MemoryReservation
• MemoryUtilization
• GPUReservation

apiGateway

• 4XXError
• 5XXError
• CacheHitCount
• CacheMissCount
• Count
• IntegrationLatency
• Latency

cloudfront

• 4XXErrorRate
• 5XXErrorRate
• 401ErrorRate
• 403ErrorRate
• 404ErrorRate
• 502ErrorRate
• 503ErrorRate
• 504ErrorRate
• BytesDownloaded
• BytesUploaded
• CacheHitRate
• OriginLatency
• Requests
• TotalErrorRate

rds

• BinLogDiskUsage
• BurstBalance
• CPUUtilization
• CPUCreditUsage
• CPUCreditBalance
• DatabaseConnections
• DiskQueueDepth
• FailedSQLServerAgentJobsCount
• FreeableMemory
• FreeStorageSpace
• MaximumUsedTransactionIDs
• NetworkReceiveThroughput
• NetworkTransmitThroughput
• OldestReplicationSlotLag
• ReadIOPS
• ReadLatency
• ReadThroughput
• ReplicaLag
• ReplicationSlotDiskUsage
• SwapUsage
• TransactionLogsDiskUsage
• TransactionLogsGeneration
• WriteIOPS
• WriteLatency
• WriteThrougput

eks

• cluster_failed_node_count
• cluster_node_count
• namespace_number_of_running_pods
• node_cpu_limit
• node_cpu_reserved_capacity
• node_cpu_usage_total
• node_cpu_utilization
• node_filesystem_utilization
• node_memory_limit
• node_memory_reserved_capacity
• node_memory_utilization
• node_memory_working_set
• node_network_total_bytes
• node_number_of_running_containers
• node_number_of_running_pods
• pod_cpu_reserved_capacity
• pod_cpu_utilization
• pod_cpu_utilization_over_pod_limit
• pod_memory_reserved_capacity
• pod_memory_utilization
• pod_memory_utilization_over_pod_limit
• pod_number_of_container_restarts
• pod_network_rx_bytes
• pod_network_tx_bytes
• service_number_of_running_pods

appSyncApi

• 4XXError
• 5XXError
• Latency
• ConnectSuccess
• ConnectClientError
• ConnectServerError
• DisconnectSuccess
• DisconnectClientError
• DisconnectServerError
• SubscribeSuccess
• SubscribeClientError
• SubscribeServerError
• UnsubscribeSuccess
• UnsubscribeClientError
• UnsubscribeServerError
• PublishDataMessageSuccess
• PublishDataMessageClientError
• PublishDataMessageServerError
• PublishDataMessageSize
• ActiveConnection
• ActiveSubscription
• ConnectionDuration

sqs

• ApproximateAgeOfOldestMessage
• ApproximateNumberOfMessagesDelayed
• ApproximateNumberOfMessagesNotVisible
• ApproximateNumberOfMessagesVisible
• NumberOfEmptyReceives
• NumberOfMessagesDeleted
• NumberOfMessagesReceived
• NumberOfMessagesSent
• SentMessageSize