in_process_exporter_metrics backtraces if process dies during scrape

[PromyLOPh]

Bug Report

Describe the bug

The process exporter input plugin dies when scraping alot of processes.

[2024/11/04 11:36:21] [engine] caught signal (SIGSEGV)
#0  0x7fc835eeefe6      in  __GI___readdir64() at inux/readdir64.c:37
#1  0x55fa41c295f8      in  process_proc_fds() at plugins/in_process_exporter_metrics/pe_process.c:767
#2  0x55fa41c2abcc      in  process_update() at plugins/in_process_exporter_metrics/pe_process.c:1131
#3  0x55fa41c2b059      in  pe_process_update() at plugins/in_process_exporter_metrics/pe_process.c:1181
#4  0x55fa41c1f14c      in  update_metrics() at plugins/in_process_exporter_metrics/pe.c:33
#5  0x55fa41c1f17e      in  cb_pe_collect() at plugins/in_process_exporter_metrics/pe.c:46
#6  0x55fa41a572a9      in  input_collector_fd() at src/flb_input_thread.c:166
#7  0x55fa41a57c2e      in  engine_handle_event() at src/flb_input_thread.c:181
#8  0x55fa41a57c2e      in  input_thread() at src/flb_input_thread.c:410
#9  0x55fa41ab4d8e      in  step_callback() at src/flb_worker.c:43
#10 0x7fc835ea8143      in  start_thread() at c:442
#11 0x7fc835f287db      in  clone3() at inux/x86_64/clone3.S:81
#12 0xffffffffffffffff  in  ???() at ???:0

To Reproduce

• Run fluent-bit -p scrape_interval=1 -i process_exporter_metrics -o stdout.
• Run while true; do for i in $(seq 10000); do sleep 5 & done; sleep 5; done at the same time.
• Observe fluent-bit crashing.

Expected behavior

No crash.

Screenshots

n/a

Your Environment

• Version used: commit 104aceb
• Configuration: (none)
• Environment name and version (e.g. Kubernetes? What version?): (none)
• Server type and version: (none)
• Operating System and version: Debian 12
• Filters and plugins: process_exporter and stdout

Additional context

Discovered while debugging #9547