Skip zlib metrics in Prometheus exporter when built with libslz

[Copilot]

When HAProxy is built with libslz instead of zlib, the Prometheus exporter exports haproxy_process_current_zlib_memory and haproxy_process_max_zlib_memory metrics with NaN values, as these metrics are only populated when USE_ZLIB is defined.

Changes

• Added preprocessor guard in promex_dump_global_metrics() to skip ST_I_INF_ZLIB_MEM_USAGE and ST_I_INF_MAX_ZLIB_MEM_USAGE when USE_ZLIB is not defined
• Metrics are now conditionally exported based on actual compression backend in use

#ifndef USE_ZLIB
    case ST_I_INF_ZLIB_MEM_USAGE:
    case ST_I_INF_MAX_ZLIB_MEM_USAGE:
        /* Skip zlib metrics when built with libslz */
        continue;
#endif

This aligns the Prometheus exporter behavior with the stats subsystem, which already conditionally populates these fields under #ifdef USE_ZLIB.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>improve prometheus metrics</issue_title>
<issue_description>when haproxy is built with libslz (not zlib), the following metrics are irrelevant

# HELP haproxy_process_current_zlib_memory Amount of memory currently used by HTTP compression on the current worker process (in bytes)
# TYPE haproxy_process_current_zlib_memory gauge
haproxy_process_current_zlib_memory NaN
# HELP haproxy_process_max_zlib_memory Limit on the amount of memory used by HTTP compression above which it is automatically disabled (in bytes, see global.maxzlibmem)
# TYPE haproxy_process_max_zlib_memory gauge
haproxy_process_max_zlib_memory NaN

we need not to add that metric if haproxy is built with slz</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes improve prometheus metrics #21

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.