Closed wreidlinger closed 1 month ago
👋 @wreidlinger
Do you use any plugins (e.g. WordPress)? Do you know what made that GET /api/plugins/v1/capabilities request?
Off-topic, but note that TZ env vars are largely unnecessary in Plausible (tz is per site) and it might cause some problems when interpreting "naive" timestamps.
I'm not suggesting changing it, but just pointing it out as something to keep in mind if you ever encounter time-related problems :)
Off-topic, but note that TZ env vars are largely unnecessary in Plausible (tz is per site) and it might cause some problems when interpreting "naive" timestamps.
I'm not suggesting changing it, but just pointing it out as something to keep in mind if you ever encounter time-related problems :)
@ruslandoga thanks for the hint. I have removed the TZ environment variables from the docker compose config. But the problem and the error still persists.
TZ env var change was not supposed to fix the error :) Please see my first comment.
Whatever is making the request is not accepting JSON as the response (it doesn't have accept: application/json
header). I'll check if the error message could be improved. And note that GET /api/plugins/v1/capabilities
is mostly a health-check endpoint, it doesn't affect visitor tracking (that goes into POST /api/event
)
👋 @wreidlinger
Do you use any plugins (e.g. WordPress)? Do you know what made that GET /api/plugins/v1/capabilities request?
Yes you are right. The domain having problems with tracking is a Wordpress site and is using the Plausible Analytics Wordpress plugin. Furthermore this Wordpress site is also running in a docker compose stack on the same virtual server as the Plausible stack is running.
Just a side info... the Plausible as well as the Wordpress stack using a Caddy webserver as there reverse proxy.
Here is the error log of the Caddy webserver for the plausible domain. To me it seams that these to error logs correlate. Looks like the internal docker IP of the Wordpress container is used for the request from the Wordpress to the Plausible stack. Strange... I'm trying to fix this issue.
{
"level": "error",
"ts": "2024/05/23 09:54:15",
"logger": "http.log.access.log5",
"msg": "handled request",
"request": {
"remote_ip": "172.26.0.5",
"remote_port": "38662",
"client_ip": "172.26.0.5",
"proto": "HTTP/1.1",
"method": "GET",
"host": "plausible.mydomain.com",
"uri": "/api/plugins/v1/capabilities",
"headers": {
"User-Agent": [
"OpenAPI-Generator/1.0.0/PHP"
],
"Accept": [
"application/json"
],
"Content-Type": [
"application/json"
],
"Authorization": []
},
"tls": {
"resumed": false,
"version": 772,
"cipher_suite": 4865,
"proto": "http/1.1",
"server_name": "plausible.mydomain.com"
}
},
"bytes_read": 0,
"user_id": "",
"duration": 0.00428649,
"size": 0,
"status": 500,
"resp_headers": {
"Server": [
"Caddy"
],
"Alt-Svc": [
"h3=\":443\"; ma=2592000"
],
"Content-Length": [
"0"
]
}
}
Hi there,
we run the plausible community edition via docker compose. Looks like not all visitors get tracked correctly. We saw the following logs, is this normal or can somebody tell me if there is a problem with the config?!