QA (public) » openQA Project (public) » openQA Infrastructure (public)

Motivation¶

We have a diverse infrastructure including hosts in multiple network segments including IPSec or wireguard tunnels between hosts which can cause problems with MTU sizes like recently in #178576. Within osado we already have a MTU size ping check checking for increasing size of packets. We can improve our salt based monitoring using
https://github.com/influxdata/telegraf/tree/master/plugins/inputs/ping but with different or higher ping sizes.

Acceptance criteria¶

• AC1: OSD monitoring provides ping monitoring for host-specific maximum MTU size
• AC2: We still have generic minimum-size ping checks in parallel
• AC3: Additional monitoring does not have a negative performance or storage overhead on any of our salt controlled hosts

Suggestions¶

• We could start with a simple maximum size ping based on what is applicable for certain network segments, e.g. based on wireguard and/or location. See https://gitlab.suse.de/openqa/salt-states-openqa/-/blob/308922361e546d6dd1d99f991cce28089d39477a/top.sls#L21 where we use 'G@needs_wireguard:True or ( *.nue2.suse.org and not G@needs_wireguard:False )'
• Add one or multiple ping checks with lower interval than default. Something like once an hour with multiple sizes should be enough
• Create an according grafana panel and alert definition
• Ensure that the additional monitoring does not have a negative performance or storage overhead on any of our salt controlled hosts