yadunut/clan-core
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

telegraf: move telegraf.json into runtime directory

Browse Source 
No need to persist this type of data. We also don't need to allocate a
persistent user for telegraf.

Also make it a directory because otherwise thre is a potential race
condition on startup.
This commit is contained in:
Jörg Thalheim
2025-09-18 08:26:07 +02:00
parent 92e2c841e3
commit 4b549365f7
3 changed files with 13 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
clanServices/monitoring/telegraf.nix
View File

@@ -56,7 +56,7 @@
enable = true;
wantedBy = [ "multi-user.target" ];
after = [ "telegraf.service" ];
wants = [ "telegraf.service" ];
requires = [ "telegraf.service" ];
serviceConfig = {
LoadCredential = [
"auth_file_path:${config.clan.core.vars.generators.telegraf.files.miniserve-auth.path}"
@@ -67,13 +67,9 @@
Restart = "on-failure";
User = "telegraf";
Group = "telegraf";
RuntimeDirectory = "telegraf-www";
};
script = "${pkgs.miniserve}/bin/miniserve -p 9990 /var/lib/telegraf/telegraf.json --auth-file \"$AUTH_FILE_PATH\"";
};
users.users.telegraf = {
home = "/var/lib/telegraf";
createHome = true;
script = "${pkgs.miniserve}/bin/miniserve -p 9990 /run/telegraf-www --auth-file \"$AUTH_FILE_PATH\"";
};
services.telegraf = {
@@ -119,7 +115,7 @@
};
outputs.file = {
files = [ "/var/lib/telegraf/telegraf.json" ];
files = [ "/run/telegraf-www/telegraf.json" ];
data_format = "json";
json_timestamp_units = "1s";
};

5
clanServices/monitoring/tests/vm/default.nix
View File

@@ -56,12 +56,12 @@
peer1.wait_for_unit("telegraf.service")
peer1.wait_for_unit("telegraf-json.service")
peer1.succeed("curl http://localhost:9990")
peer1.succeed("curl http://localhost:9990/telegraf.json")
peer1.succeed("curl http://localhost:9273/metrics")
# Fetch the basic auth password from the secret file
password = peer1.succeed("cat /var/run/secrets/vars/telegraf/password")
url = f"http://192.168.1.1:9990"
url = f"http://192.168.1.1:9990/telegraf.json"
credentials = f"prometheus:{password}"
print("Using credentials:", credentials)
time.sleep(10) # wait a bit for telegraf to collect some data
@@ -81,7 +81,6 @@
found_system = True
print("Found nixos_systems metric in json output")
break
print(line)
assert found_system, "nixos_systems metric not found in json output"
# TODO: I would like to test the python code here but it's not working yet

2
pkgs/clan-cli/clan_lib/metrics/telegraf.py
View File

@@ -36,7 +36,7 @@ def get_metrics(
"""
# Example: fetch Prometheus metrics with basic auth
url = f"http://{target_host.address}:9990"
url = f"http://{target_host.address}:9990/telegraf.json"
username = "prometheus"
var_name = "telegraf/password"
password_var = get_machine_var(machine, var_name)