AC:
- Create a working dashboard for the Cirrus Streaming Updater. A first attempt is here .
- Related dashboards:
- Flink-App : should be able to use as-is
- WDQS Streaming Updater Can use producer and consumer lag
- Link relevant dashboards together
- Create alerts
- Decide alert urgency and paging strategy. As @dcausse pointed out today, the downstream services like ChangeProp are closely watched by mainline SREs, as they are used by a lot more than just search functionality. When the Search Update Pipeline stops using ChangeProp and starts using Flink, we lose mainline SRE visibility. They also don't have experience with Flink, so help will be limited. That means we'll have to watch closer and react more quickly.
- Link all dashboards in cirrus-streaming-updater documentation.