chore(infra): drop redpanda metrics scrape source from Vector

Redpanda Cloud doesn't expose /public_metrics the way the self-hosted
Hetzner box did, so the prometheus_scrape source had been failing on
every scrape since the migration. Drop the source, the transform that
tagged it, and the Axiom sink that consumed it.

Cluster-level metrics are now visible in the Redpanda Cloud dashboard.

Author: izadoesdev

infra/ingest/vector.yaml

@@ -24,13 +24,6 @@ sources:
     decoding:
       codec: json
 
-  redpanda_metrics:
-    type: prometheus_scrape
-    endpoints:
-      - "${REDPANDA_METRICS_ENDPOINT}"
-    scrape_interval_secs: 15
-    scrape_timeout_secs: 5
-
 transforms:
   route_analytics:
     type: route
@@ -49,14 +42,6 @@ transforms:
       uptime_checks: '.topic == "analytics-uptime-checks"'
       link_visits: '.topic == "analytics-link-visits"'
 
-  redpanda_metrics_transform:
-    type: remap
-    inputs:
-      - redpanda_metrics
-    source: |
-      .tags.cluster = "${REDPANDA_CLUSTER_NAME}"
-      .tags.env = "${ENVIRONMENT}"
-
 sinks:
   clickhouse_events:
     type: clickhouse
@@ -285,11 +270,3 @@ sinks:
     path: /var/log/vector/dead_letters.log
     encoding:
       codec: json
-
-  axiom_redpanda_metrics:
-    type: axiom
-    inputs:
-      - redpanda_metrics_transform
-    dataset: "${AXIOM_DATASET}"
-    token: "${AXIOM_TOKEN}"
-    compression: gzip