Globally evaluate rules for services using Runway
Globally evaluate rules for services using Runway
These services are using a Prometheus to which we don't deploy any recording rules. Instead Prometheus uses thanos-receive to get all of the metrics scraped from the running service into Thanos.
This means that we need to evaluate and deploy the recording rules there.
This applies this to the Runway-archetype, and as a result we modify
the recording rule definitions for the ai-gateway
service.
The ai-gateway uses old openmetrics histograms
The code_suggestions
one was using 10.0
but the ai-gateway
is
using 10
.
This is likely due to different Prometheus client versions. The
code_suggestions
service is using the newer version, so it is likely
to come back in the future. That is why we're setting the state to
"migrating" for the histograms in the ai-gateway
service.