Description of problem:
Current SELinux policy is preventing the Performance Metrics Domain Agent (PMDA) for nginx (part of the Performance Co-Pilot (PCP) suite) access to nginx's statistics available at http://localhost/nginx_status.


Version-Release number of selected components:
selinux-policy-targeted-3.13.1-128.12.fc22.noarch
pcp-pmda-nginx-3.10.6-1.fc22.x86_64
nginx-1.8.0-10.fc22.x86_64


How reproducible:
Always.

Steps to Reproduce:
1. sudo dnf install nginx
2. create a file in /etc/nginx/default.d/nginx_status.conf with the following contents:
location /nginx_status {
    stub_status on;
    access_log off;
    allow 127.0.0.1;
    allow ::1;
    deny all;
}
3. sudo systemctl restart nginx.service
4. sudo dnf install "perl(LWP::UserAgent)" pcp-collector
5. sudo touch /var/lib/pcp/pmdas/nginx/.NeedInstall
6. sudo systemctl restart pmcd.service
7. pmval nginx.requests_count

Actual results:
[root@collector vagrant]# pmval nginx.requests_count 

metric:    nginx.requests_count
host:      collector
semantics: cumulative counter (converting to rate)
units:     count (converting to count / sec)
samples:   all

pmval: pmFetch: Unknown or illegal metric identifier

pmval: pmFetch: Unknown or illegal metric identifier

pmval: pmFetch: Unknown or illegal metric identifier

pmval: pmFetch: Unknown or illegal metric identifier

The relevant line from /var/log/audit/audit.log:
type=AVC msg=audit(1442242235.474:225): avc:  denied  { name_connect } for  pid=532 comm="perl" dest=80 scontext=system_u:system_r:pcp_pmcd_t:s0 tcontext=system_u:object_r:http_port_t:s0 tclass=tcp_socket permissive=0


Expected results:
[root@collector vagrant]# pmval nginx.requests_count 

metric:    nginx.requests_count
host:      collector
semantics: cumulative counter (converting to rate)
units:     count (converting to count / sec)
samples:   all
               0.9982
               0.9982
               0.9981