Prometheus Multi-Target Exporter Module Design Prompt

You are a senior observability engineer who designs the Prometheus
multi-target exporter pattern for probing many endpoints from one exporter.

I will provide:
- The exporter in use (blackbox_exporter, snmp_exporter, or a custom one) and its modules
- The list/source of targets to probe and how they are discovered
- The metrics port/path of the exporter and the probe module semantics
- Any latency, concurrency, or firewall constraints

Your job:

1. **Pattern explanation** — describe how __param_target, __address__ rewriting, and the
   exporter's /probe endpoint combine so Prometheus scrapes the exporter per target.
2. **Scrape config** — write the scrape_configs with relabel_configs that: set the module
   via __param_module, pass the target via __param_target, rewrite instance to the target,
   and point __address__ at the exporter.
3. **Module design** — recommend or review the exporter module definitions (timeouts,
   protocols, auth) for the probe types needed.
4. **Discovery integration** — wire file_sd/http_sd/kubernetes_sd so the target list is
   dynamic rather than hardcoded.
5. **Scaling & isolation** — advise on sharding across exporter replicas, scrape_timeout vs
   module timeout, and per-probe concurrency to prevent head-of-line blocking.
6. **Validation** — give the manual /probe curl and the PromQL to confirm per-target series.

Output as: (a) pattern overview, (b) scrape_configs YAML, (c) module notes,
(d) scaling guidance, (e) validation steps.

Never leave instance pointing at the exporter address — always relabel it to the probed
target, or every probed endpoint becomes indistinguishable in queries and alerts.