Problem/Question
I recently updated my Netdata Parent and Children nodes to v2 and I’m unable to export netdata parent and child metrics to influxdb using telegraf. I point telegraf to get data from
the Netdata API endpoint: http://parent-node-netdata:19999/api/v1/allmetrics?format=prometheus_all_hosts , but the scraping fails due to incorrect metrics format. The issue ie that the metrics format contains repeated HELP and TYPE for same metrics, which is NOT as per Prometheus format.
This works in the 1.45 version, but not in versions 1.46 and later.
Relevant docs you followed/actions you took to solve the issue
-
I referred to Netdata’s Release page Releases · netdata/netdata · GitHub and tested it on various releases: 1.45.5, 1.45.6, 1.46.0,1.46.1,1.46.2,1.46.3,1.47.4.
It works for 1.45, and not on versions that came later. -
Related Netdata MRs that has done work on this:
Compatibility with Prometheus HELP by thiagoftsm · Pull Request #17191 · netdata/netdata · GitHub
fix prometeus export: missing comma before "instance" label by ilyam8 · Pull Request #18061 · netdata/netdata · GitHub -
Related Prometheus error: Input.prometheus plugin unable to parse #HELP comment. · Issue #8366 · influxdata/telegraf · GitHub
Environment/Browser/Agent’s version etc
- Parent Node:
Netdata version: v2.0.3
Type:
docker
O/S version:Ubuntu, 22.04.4 LTS (Jammy Jellyfish)
Architecture:x86_64
Kernel:Linux, 6.8.0-1018-gcp
CPU:2.20 GHz (2 Cores)
Memory:11.68 GiB RAM
Hard disk size:100.00 GiB
- Children Nodes:
Netdata version: v2.0.3
Type:
lxc
O/S version:Ubuntu, 22.04.4 LTS (Jammy Jellyfish)
Architecture:aarch64
Kernel:Linux, 5.15.136-tegra
CPU:1.98 GHz (8 Cores)
Memory:15.29 GiB RAM
Hard disk size:238.47 GiB
What I expected to happen
I expected the Telegraf would export metrcis to Influxdb.