Most hosts will not need to send any messages to ntfy. Let's define the ntfy pipeline stages only for the machines that need them. There are currently two use cases for ntfy: * MD RAID status messages (from Chromie and nvr2) * WAN Link status messages (from gw1) Breaking up the pipeline into smaller pieces allows both of these use cases to define their appropriate filters while still sharing the common steps. The other machines that have no use for these steps now omit them entirely.
21 lines
413 B
YAML
21 lines
413 B
YAML
fluent_bit_ntfy_filters:
|
|
- name: rewrite_tag
|
|
alias: ntfy
|
|
match: host.*
|
|
rule: transport kernel ntfy true
|
|
- name: grep
|
|
match: ntfy
|
|
alias: ntfy.filter
|
|
regex: message ^md
|
|
|
|
fluent_bit_default_filters: >-
|
|
{{
|
|
fluent_bit_common_filters +
|
|
fluent_bit_ntfy_filters +
|
|
fluent_bit_ntfy_common_filters
|
|
}}
|
|
|
|
fluent_bit_main_outputs:
|
|
- '{{ fluent_bit_output_systemd }}'
|
|
- '{{ fluent_bit_output_ntfy }}'
|