Logging Beijing - M3 Architecture Review

Owned by Michael O'Brien
Last updated: May 10, 2018
1 min read

Copied template from the AAI project - Thank @James Forsyth

Delta from R1

Started the library (for cassablanca)

https://lf-onap.atlassian.net/browse/LOG-118

Deliver python library https://lf-onap.atlassian.net/browse/MULTICLOUD-151 in the logging repo via the Multicloud team (@Ramki Krishnan, @Shankaranarayanan Puzhavakath Narayanan, and @Sastry Isukapalli interested). Note: @Sastry Isukapalli has an implementation based on ECOMP's EELF for Python and can share notes and code from that.

Sonar must be above 50%, and address CLM



Future Cassablanca

  1. We retrofitted the project as a Java project (jenkins, sonar)

  2. We are developing a Java RI - planning completion for Cassablanca
    https://lf-onap.atlassian.net/browse/LOG-95



  3. The Multicloud team has committed a python RI under

  4. The Multicloud team is retrofitting our repo for python

  5. We are looking at alternatives to the sidecar for the filebeat container

  6. we are looking at out of the box transaction tracing libraries like opentrace

  7. we are looking to investigate logging in DCAEGEN2 - first step was to bring up the full DCAE stack

S3P Updates

  1. Security

    1. Logging RI:  

      98% for 

      https://bestpractices.coreinfrastructure.org/projects/1578

       

    2. SONAR code coverage.  Plan is to get to >50% on all repos.

    3. TODO: communicate with security team

    4. Nexus IQ scans: 

  2. Scalability and Resiliency

    1. Working closely with OOM team (all Logging developers are also OOM contributors)

    2. Relying on kubernetes to manage Logging ELK stack resilience

  3. Performance and stability

    1. Focus to this point has been security and scalability/resiliency.  We will participate in the integration team's performance testing

    2. ELK stack can meet soak test requirements

  4. Manageablility

    1. Logging Filebeat sidecar refactor
      Possible sidecar refactor to use a deamonset in 

    2. ELK stack can be instantiated in < 1hr - around 5-10 min as part of overall OOM deployment

  5. Usability

Deployment Configuration Alignment

  1. We are aligning with the centralized configuration work in OOM








API Updates

We currently have no logging library for Beijing - the stub work is in