Introduce option for high volume request tracking

bpf/k_tracer.c

@@ -407,6 +407,9 @@ int BPF_KRETPROBE(kretprobe_tcp_sendmsg, int sent_len) {
 }
 
 static __always_inline void ensure_sent_event(u64 id, u64 *sock_p) {
+    if (high_request_volume) {
+        return;
+    }
     send_args_t *s_args = bpf_map_lookup_elem(&active_send_args, &id);
     if (s_args) {
         bpf_dbg_printk("Checking if we need to finish the request per thread id");

bpf/protocol_http.h

@@ -147,6 +147,9 @@ get_or_set_http_info(http_info_t *info, pid_connection_info_t *pid_conn, u8 pack
 }
 
 static __always_inline void finish_possible_delayed_http_request(pid_connection_info_t *pid_conn) {
+    if (high_request_volume) {
+        return;
+    }
     http_info_t *info = bpf_map_lookup_elem(&ongoing_http, pid_conn);
     if (info) {
         finish_http(info, pid_conn);

charts/beyla/README.md

@@ -24,7 +24,7 @@ eBPF-based autoinstrumentation HTTP, HTTP2 and gRPC services, as well as network
 |-----|------|---------|-------------|
 | affinity | object | `{}` | used for scheduling of pods based on affinity rules |
 | config.create | bool | `true` | set to true, to use the below default configurations |
-| config.data | object | `{"attributes":{"kubernetes":{"enable":true},"select":{"beyla_network_flow_bytes":{"include":["k8s.src.owner.type","k8s.dst.owner.type","direction"]}}},"filter":{"network":{"k8s_dst_owner_name":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"},"k8s_src_owner_name":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"}}},"prometheus_export":{"path":"/metrics","port":9090}}` | default value of beyla configuration |
+| config.data | object | `{"attributes":{"kubernetes":{"enable":true},"select":{"beyla_network_flow_bytes":{"include":["k8s.src.owner.type","k8s.dst.owner.type","direction"]}}},"filter":{"application":{"client":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"},"server":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"}},"network":{"k8s_dst_owner_name":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"},"k8s_src_owner_name":{"not_match":"{kube*,*jaeger-agent*,*prometheus*,*promtail*,*grafana-agent*}"}}},"prometheus_export":{"path":"/metrics","port":9090}}` | default value of beyla configuration |
 | config.name | string | `""` |  |
 | dnsPolicy | string | `"ClusterFirstWithHostNet"` | Determines how DNS resolution is handled for that pod. If `.Values.preset` is set to `network` or `.Values.config.data.network` is enabled, Beyla requires `hostNetwork` access, causing cluster service DNS resolution to fail. It is recommended not to change this if Beyla sends traces and metrics to Grafana components via k8s service. |
 | env | object | `{}` | extra environment variables |