The examples in this folder demonstrate various elements of a more complicated daemonset and aggregation tier deployment.
- bootstrap_ds.yaml: This is the bootstrap configuration for the daemonset. It writes meta stats to the agg tier, and identifies itself via the Kubernetes node and pod names (delivered via the upward API). The actual configuration is loaded from a config map which can be hot reloaded.
- bootstrap_agg.yaml: This is the bootstrap configuration for the aggregation tier. It writes meta stats to an APS workspace, using default AWS auth. The actual configuration is loaded from a config map which can be hot reloaded.
Both bootstrap configurations expose the admin API on port 9999.
- config_ds.yaml: This configuration demonstrates a pipeline that scrapes Prometheus metrics from various sources. It does various mutations before sending them off to the aggregation tier.
- config_agg.yaml: This configuration demonstrates a pipeline that interoperates with a control plane to send metric samples and receive remote controlled block lists. It also performs aggregation on incoming metrics from the daemonsets. Finally the metrics are sent off to an APS workspace via default AWS auth. The aggregation configuration is tuned for Prometheus, and a decision has been made to convert absolute counters to delta counters (pre-rate) and to send stale markers. This makes the entire setup work better with zero elision and blocking.