ความคิดเห็น •

  • @GM-qv1ql
    @GM-qv1ql 3 ปีที่แล้ว

    Nice and simple explanation!

  • @TomerBenDavid
    @TomerBenDavid 4 ปีที่แล้ว +1

    Yeah!!!! Loved the best practice section.

  • @umaradhami
    @umaradhami 4 ปีที่แล้ว +1

    Believe me !!!! You have the best content

  • @Roberto-live
    @Roberto-live 3 ปีที่แล้ว

    Great info!!

  • @anandvamsi5141
    @anandvamsi5141 4 ปีที่แล้ว

    Very informative

    • @BeABetterDev
      @BeABetterDev 4 ปีที่แล้ว

      Glad you liked it Anand!

  • @orien758
    @orien758 8 หลายเดือนก่อน

    Great explained. AI use it's awesome.

  • @vidyasankar4661
    @vidyasankar4661 3 ปีที่แล้ว

    Hey , Good and apt content. Can u provide me the blog link. Iam looking for more possibilities than what you have given in video. Can u provide more use cases with anomaly detection .Iam working on an anomaly detection and looking for more possibilities.

  • @aradhnasingh2157
    @aradhnasingh2157 3 ปีที่แล้ว +2

    Thanks for the great content.
    I have a question.How to add this metric -(Like Order per minute) to CloudWatch? I want to apply anomaly detection on my sensor data . Could this anomaly detection work on taking a sensor reading(e.g-Temperature) and then identifying the anomaly in that univariate time series.The sensor readings are coming in redshift.

    • @BeABetterDev
      @BeABetterDev 3 ปีที่แล้ว

      Hi Aradhna,
      Good question. If your data is sourced in redshift, what you could do is soemthing like the following:
      1. Create Cloudwatch event that ticks every 1 minute.
      2. Create a lambda function that queries on your sensor data from Redshift. In this lambda, emit metrics to cloudwatch for your sensor temperature.
      3. Connect your cloudwatch event to invoke lambda
      Using this approach, you'll be able to collect 1min entries in cloudwatch and set up anomaly detection!

    • @aradhnasingh2157
      @aradhnasingh2157 3 ปีที่แล้ว

      @@BeABetterDev Thank You so much for such clearty ,following which I created a lambda and created "put metric" in CloudWatch and added a schedule of 1 min.
      Just one thing more ,do we need to keep on collecting that 1 min entries for certain period so that when we switch on anomaly detection the CloudWatch event take the past data and then create bands of upper and lower limits?Does that's how machine learning internally detects pattern? Or should I start on the anomaly detection from the day the CloudWatch start collecting the 1 min data.

  • @yunli2742
    @yunli2742 3 ปีที่แล้ว

    Hey how did you know the cost of anomaly detection?

  • @yurilppm
    @yurilppm 3 ปีที่แล้ว

    Do you think Anomaly Detection is recomended for billing alarms? Great video congrats!

    • @BeABetterDev
      @BeABetterDev 3 ปีที่แล้ว

      Hi Yuri!
      Using ML on billing is definitely a great idea! Let me know if you get this working!

  • @saikiranreviews
    @saikiranreviews 2 ปีที่แล้ว +1

    If I add, anamoly detection in a widget for a graph will it cost me extra?

    • @BeABetterDev
      @BeABetterDev 2 ปีที่แล้ว

      Hi Saikiran,
      If you go to the below URL and click the "Alarms" tab, you can see the pricing structure for anomaly detection. It does cost extra to use this feature.
      aws.amazon.com/cloudwatch/pricing/
      Hope this helps,
      Daniel