1. Preparation (e.g. previous action points)

1. Sander finishes the documentation (identifies gaps in current documentation, separates old from new information, centralizes documentation to SystemAdministration/Monitoring/Icinga).
2. Everyone reads the documentation no sooner than three working days before the VC, discusses it internally (if needed, e.g. ASV), and comes up with a list of questions and suggestions (if needed).
3. Everyone checks whether logging in via SSH to the monitoring host works for them, whether they have the desired permissions, and whether they can work with the system (OS, tools, configuration, filesystem layout, etc.).
4. Everyone who wants write access to ​the GitHub repo with the Icinga config but does not have it yet submits his GitHub account name to ​monitoring@clarin.eu. Sander fulfills these requests.

2. Agenda

The VC will be held via ​Adobe Connect.

The following descriptions have the status of notes, and may be amended/elaborated by anyone involved.

2.1. Documentation walkthrough

We walk through the documentation, and Sander resolves any issues live. This makes sure that they are addressed immediately without back-and-forth over e-mail, and satisfactorily. Only if issues are very major and hard to resolve at once, will the task be postponed, but then Trac tickets should be made to specify and track each such task.

2.2. Workflow

Benedikt will clarify the workflow insofar needed, including the GitHub <-> monitoring host configuration management script, described here ​https://github.com/clarin-eric/monitoring/blob/master/README.md.

2.3. Shibboleth SP configuration

The DNS name monitoring.clarin.eu is now in place for the monitoring host. The SAML metadata about the Shibboleth SP needs to be updated. A valid TLS certificate needs to be put in place. We will discuss the issue insofar needed, and assign tasks.

2.4. Discuss Icinga 1.x -> 2.x migration

1. Changes
   • Changing configuration syntax:
     • both not json or any standard, so there are no advantages of the new syntax (?)

       old:
       define host {
           use             pnp-host
           host_name       clarin.fz-juelich.de
           address         134.94.199.71
       }
       new:
       object Host "clarin.fz-juelich.de" {
               import "generic-host"
               address = "134.94.199.71"
       }
       

   • Are there more plugins available?
   • Is it possible to accept incoming NSCA results? (Leipzig?) At a first glance, yes (Juelich).

2. Risks?
   • The configuration generation script that is querying the Centre Registry must be reimplemented
     • Pynag is not available for the management of Icinga 2 configurations, are there other solutions that are equal dynamic?

2.5. Decision: Icinga 1.x or 2.x?

Decision alternatives:

1. First attempt to migrate to Icinga 2.x, but abandon, if this is found too unpractical, for Icinga 1.x.
2. First go Icinga 1.x, after this has succeeded start work (in parallel setup) to migrate to Icinga 2.x.
3. ...

2.6. Decision: outline of time planning for migration

Who does what when, and what are the dependencies for tasks?

2.7. Decision: follow-up actions

Follow-up VC??

3. Minutes

...

4. New action points

1. Set up Let's Encrypt for certificate. (Thomas Hynek)

2. Propose changed version of SAML metadata about the Shibboleth SP. (Sander)

3. Update SP configuration (based on 2) in a way that Puppet agrees with it. (Benedikt)

Follow-up moment Thursday 6 April afternoon, Sander sends status e-mail around to ​monitoring@clarin.eu.