[che-dev] Che community call 2018-07-16: Meeting notes

[Dave Neary <dneary@xxxxxxxxxx>]

Hi all,

We had a Che community meeting, as proposed on the mailing list last
week, on our usual "off week", on the topic of logging, tracing, and
monitoring for Che (platform, core application, and workspaces).

The meeting was led by Tamir Menahem and Yariv Lifchuk of SAP. Here are
the notes from the call:

https://docs.google.com/document/d/1yF6DjW6chfuit-oHZmrGqvv2-1Y2tSN1j4YLhQPgVCM/edit#


Text notes:

Log, Trace, Monitoring

* Log epic - https://github.com/eclipse/che/issues/10290
* Trace epic - https://github.com/eclipse/che/issues/10298
* Monitor epic - https://github.com/eclipse/che/issues/10329

Yariv’s goal is to summarise recommendations, and open discussion on
this topic
 - Goal is to align the community on log format standardization
 - Logs should go to stdout/err to enable log streaming
 - Want to add mandatory params for user_id, tenant, and optional params
which could be useful for some clients (eg trace_id)
 - Recommends using Prometheus for monitoring guest instances. Covers
system monitoring. For application monitoring, a specific health check
is necessary
 - Want to use logs to send monitoring events to central repository with
a custom tag
 - Need both internal app status monitor/health check, and an external
health check to monitor whether instances stop responding
 - With all these health checks, proposal to use a healthcheck manager
to manage all of the healthchecks

SAP prioritization:
1. System monitoring - monitoring k8s, Che services
2. Logging: fluentd or logstash for log aggregation
3. Critical external health checks: add health check end-point to
workspace pods

There was a long and detailed conversation about restart policy and
severity of events for containers in pods.

Mario and Stevan propose a separate call to track progress for logging
and system monitoring.

System monitoring: Expecting work to complete in next 1-2 sprints
Logging: Another ~3 sprints after this

Tamir: To set up another meeting for follow-up on logging, to include
Mario, Sergii, and Alex (and anyone else interested) for follow-up call

Thanks,
Dave.


On 07/09/2018 06:25 PM, Dave Neary wrote:
> Hi everyone,
> 
> A number of key participants in the Che project will be unavailable in 2
> weeks for the next scheduled Che meeting. I would like to propose moving
> the meeting up a week, and having it next Monday, at the usual time.
> 
> Does anyone have any objections? The first topic will be on tracing and
> metrics, Tamir, will you be available to discuss this next week?
> 
> Thanks,
> Dave.
> 

-- 
Dave Neary - Open Source and Standards team
Red Hat - http://community.redhat.com
Ph: +1-978-399-2182 / Cell: +1-978-799-3338