A Service Weaver application has logs, traces, and metrics. Before this PR, weaver kube deploy generated Kubernetes YAML to set up Prometheus, Jaegar, Grafana, and Loki. The weaver kube babysitter exported logs, traces, and metrics to these systems. However, there was no way for a user to customize how telemetry was handled. If you wanted to export traces to Datadog, for example, well, you couldn't.
To address this problem in the long term, we're considering introducing a general purpose mechanism that would allow developers to customize deployers using a set of components. A metric exporting component, for example, would customize how metrics are exported.
In the short term, we're going to introduce a much simpler and less general approach where a developer can plug in custom ways to manage telemetry. This also means that weaver kube no longer has to generate YAML to run Prometheus, Jaegar, and whatnot. Instead, they will be plugins.
A Service Weaver application has logs, traces, and metrics. Before this PR,
weaver kube deploygenerated Kubernetes YAML to set up Prometheus, Jaegar, Grafana, and Loki. Theweaver kubebabysitter exported logs, traces, and metrics to these systems. However, there was no way for a user to customize how telemetry was handled. If you wanted to export traces to Datadog, for example, well, you couldn't.To address this problem in the long term, we're considering introducing a general purpose mechanism that would allow developers to customize deployers using a set of components. A metric exporting component, for example, would customize how metrics are exported.
In the short term, we're going to introduce a much simpler and less general approach where a developer can plug in custom ways to manage telemetry. This also means that
weaver kubeno longer has to generate YAML to run Prometheus, Jaegar, and whatnot. Instead, they will be plugins.