...
- Review guidelines in general
- Review line formats
Comparison of line formats: from Luke
Before:
Code Block language text 11/08-02:11:28.213|org.onap.example.component1.subcomponent1.LogbackTest|b88319fb-966e-4eb9-ad7f-a604da21c55a||main|||INFO||||||here's an info it has newlines and tabs
After
Code Block org.onap.example.component1.subcomponent1.LogbackTest 2017-11-07T15:11:28.213Z INFO here's an info\nit has newlines\tand tabs key1=value1, key2=value2 with space, RequestId=b88319fb-966e-4eb9-ad7f-a604da21c55a, key5=value5"with"quotes, key6=value5>with<chevrons, key3=value3\nwith\nnewlines, key4=value4\twith\ttabs main
- TODO: need to correlate the position of the thread name "main" in its new position in the guidelines, also the MDC comma separated block. We need to specify the exact tab structure preceding the mdc block
- Introduce 2 new documentation contributors: Akanksha Khosla (OOM) , Anuj Kapoor (LOG)
- discuss doc work - todo
- discuss doc release notice work
- discuss carrier grade work in OOM -OOM for Carrier Production-Grade Deployments around
Jira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key OOM-109 - JIRA review - https://jira.onap.org/secure/RapidBoard.jspa?rapidView=53&view=planning&selectedIssue=LOG-85
- Triage log issues in other components, AAI, OOM - introduce Logging label
,Jira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key AAI-454 Jira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key OOM-403 - Discuss overall ONAP stability as it is affecting our log traceability -
please share general SDC, MSO, AAI workarounds hereJira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key INT-106 - F2F Dec 2017 presentations - Transaction Traceability using ELK by James
Minutes
- Discussed
- Discuss delimiter, new line, mdc name/value pairs
- Discuss producers (devs) as well as consumers (ops) AT&T and Bell - need to handle ops as well as NBI clients - from an indexing/parsing perspective
- Discuss use case where indexing fails and ops reads the logs directly
- Allow for default and new format toggling - could have dual appenders/or subdirs for both formats until R2? - see MSO for example in ONAP Log Locations
- maintain backward compatabilty with EELF
- Standardize in the future on logger names not different files
- Need to rework standard directory issue
- Review log eviction policy for emptyDir and dockerdata-nfs outputs
- Check ELK container logstash HD
- Compare overhead of logging over NFS - need some volumetrics state before/after log duplication
- part of log audit - verify eelf appenders in logback.xml
Post Meeting
- Epic for dual appenders (eelf, current) design issues, Resiliency/rotation changes, migration to single appender in the future - Mike
- Epic for default configuration being applied
- Finalize tabs vs pipes for R2 doc - vote on tue
- Implementation (slf4j/eelf) is less of a concern over verifying that indexed content of the log format is standardized for consumers - however an additional API wrapper would benefit abstractONAP-component-wide changes over direct SLF4J calls..1
- Enumerate new ONAP projects as candidates for guideline enforcement.
- Need volumetrics performance stats for end to end log emission all the way to ELK indexing/query - to be able to answer for more real time requestID tracking - like a DMaaP queue?
- Committers have ideas about a 3rd appender to drop into the message queue DMaaP - a secondary, perhaps more performant view of the events.
- Design Issue Epic for above
Jira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key LOG-94 - AOP annotation support for Logging - check for existing code and raise (custom EELF annotations on top of Spring ?) -
Jira Legacy server System Jira serverId 4733707d-2057-3a0f-ae5e-4fd8aff50176 key LOG-95
Rollover Actions
From last week
...