Discourse Prometheus

AI-generated summary

The Discourse Prometheus plugin is an official exporter that collects key metrics from Discourse and exposes them at the /metrics endpoint for consumption by Prometheus. As detailed in the introduction, the plugin aggregates data from all forked unicorn processes into a single endpoint. It tracks various metrics including median and 99th percentile response times for pages, execution time breakdowns (SQL/Redis/App), page views, error tracking, Ruby object space statistics, V8 memory stats, and Sidekiq job durations. A sample Grafana dashboard is available on Grafana.com.

Access control is managed via the prometheus_trusted_ip_allowlist_regex global setting (or DISCOURSE_PROMETHEUS_TRUSTED_IP_ALLOWLIST_REGEX environment variable). pfaffman clarified that this setting must be configured in app.yml, addressing confusion from inlann. This allows users to define trusted IP addresses for the metrics endpoint, which is particularly useful for dynamic IPs.

Several users raised specific questions and issues:

  • Mudit_Maheshwari asked if metrics aggregate across multiple Kubernetes pods and questioned why only specific controller action timings are observed. (Note: The original post suggests it aggregates per instance, but cross-pod aggregation typically requires service discovery in Prometheus).
  • DevTeVe reported integration failures with Grafana, showing errors in /metrics/api/v1/query_range requests. They later found a solution in a separate thread regarding 404 errors.
  • lavamind asked if the plugin exports a metric for available Discourse upgrades. NateDhaliwal confirmed this is handled by the Docker Manager plugin, not Prometheus. Nate subsequently shared a workflow using the workflows plugin to notify users of updates.
:discourse2: Summary Discourse Prometheus is the official Prometheus exporter for Discourse
:hammer_and_wrench: Repository Link https://github.com/discourse/discourse-prometheus
:open_book: Install Guide How to install plugins in Discourse

The Discourse Prometheus plugin collects key metrics from Discourse and exposes them in the /metrics path so prometheus can consume them.

These metrics can be used to Graph all sorts of data like:

Median and 99th percentile times for topic / categories / top and latest pages. Breaking down execution time between SQL/Redis and App.

Page view tracking

Error tracking

Ruby object space tracking including allocation rate, heaps and so on.

Hosted V8 memory statistics

Scheduled Job Queue and Sidekiq job durations and executions.

To see a full list of metrics available, install the plugin and visit SITENAME/metrics as an admin.

Out of the box we allow the metrics route to admins and private ips.


Discourse Prometheus is smart enough to aggregate data for all forked unicorn processes and present it as cohesive metrics on a single endpoint. We use it internally to keep track of our sites.

Sample dashboard at:


Discourse global setting called prometheus_trusted_ip_allowlist_regex (env: DISCOURSE_PROMETHEUS_TRUSTED_IP_ALLOWLIST_REGEX) is allowing to set Prometheus’s trusted IP addresses. The setting accepts regular expression to set advanced IP ranges. This is useful in a case when Prometheus IP address is dynamic.

Last edited by @JammyDodger 2024-06-03T21:55:48Z

Check documentPerform check on document:
52 Likes
Install Prometheus plugin + Prometheus on your server
Add more metrics to Discourse Prometheus
Prometheus exporter problem with SecureRandom (multisite?)
Webhook for Discourse Uptime Monitoring?
Best configurations for speeding up standalone discourse
Huge increase in Redis use after changing hosts
Prometheus scrape job cannot reach the metrics
Prometheus pluging Cross-Origin Request
How to avoid upstream timeouts?
Staging instance on same Production instance Docker Server?
A lot of CPU cores or fast cores?
Memory creep in last couple of updates
What's the correct REGEX for the allowlist in app.yml?
Tracking discourse usage by hour?
Discourse disk space limits per user?
Admin Reporting & Analysis: Incremental Changes
Out of memory errors with custom plugin
Postgres connection error on 'bundle exec rails c'
Getting extreme load warning while server resources are not being used
Real-time updating of topics freezes under high activity
Docker image update: Redis 6 and 25% smaller image size
What happens next?
Real-time updating of topics freezes under high activity
Real-time updating of topics freezes under high activity
High load due to peak anonymous sessions, increase unicorn workers?
Opinions on AWS vs DO Page Speed / Performance
Can I ignore some user agents?
Extreme Load Error After Upgrading to 3.3.0.beta3-dev Yesterday (on Prem)
Multiple Discourse Instances with WordPress, Ghost and subfolder support
Display community health stats on lemon (lemon = RPi + Pimoroni Unicorn Hat HD)

I’m currently managing discourse on Kubernetes, running on multiple app and sidekiq pods. I have a couple of questions regarding the plugin:

  1. Does the plugin aggregate metrics from all the app pods? I’m assuming it does aggregate metrics for unicorn processes running on a single instance/pod.
  2. I’m curious about the reason behind observing timings for only some controller actions. This query is related to the code at discourse-prometheus/lib/collector.rb at main · discourse/discourse-prometheus · GitHub.

Hello,

Where can I find the setting prometheus_trusted_ip_allowlist_regex? :crying_cat_face:

I’m using Discourse v3.4.0.beta3-dev

You have to set DISCOURSE_PROMETHEUS_TRUSTED_IP_ALLOWIST_REGEX in app.yml.

1 Like

Not sure what’s happening but this integration is not working anymore. I couldn’t get the dashboard in grafana to display anything… the /metrics does return something but anything beyond that does not really work

Many errors like the one below

Started GET "/metrics/api/v1/label/instance/values?limit=40000&start=1760980620&end=1760982480&match%5B%5D=discourse_page_views" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started GET "/metrics/api/v1/label/job/values?limit=40000&start=1760980620&end=1760982480&match%5B%5D=discourse_page_views%7B%7D" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
  Rendered layout layouts/no_ember.html.erb (Duration: 21.1ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 26.0ms | GC: 0.0ms)
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
  Rendered layout layouts/no_ember.html.erb (Duration: 19.8ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 28.2ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 34.6ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 44.4ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 41.7ms | GC: 0.0ms)
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:07 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:08 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:08 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:08 +0000
  Rendered layout layouts/no_ember.html.erb (Duration: 23.6ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 24.9ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 19.9ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 21.6ms | GC: 0.0ms)
Started GET "/metrics/api/v1/label/job/values?limit=40000&start=1760980620&end=1760982480&match%5B%5D=discourse_page_views%7B%7D" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
Started GET "/metrics/api/v1/label/instance/values?limit=40000&start=1760980620&end=1760982480&match%5B%5D=discourse_page_views" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
  Rendered layout layouts/no_ember.html.erb (Duration: 19.5ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 20.9ms | GC: 0.0ms)
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
Started POST "/metrics/api/v1/query_range" for 16.63.57.20 at 2025-10-20 17:47:37 +0000
  Rendered layout layouts/no_ember.html.erb (Duration: 22.4ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 22.0ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 40.0ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 37.0ms | GC: 0.0ms)
  Rendered layout layouts/no_ember.html.erb (Duration: 24.8ms | GC: 1.6ms)

Cross-posting this as I found a solution that worked for me Discourse Prometheus 404 metrics for Grafana - #2 by DevTeVe

Does the plugin export a metric that indicates whether a Discourse upgrade is available?

That will be in the Docker Manager plugin which is installed by default, but only viewable when you are in Discourse.

But perhaps the workflows plugin has a way of notifiying a user if there is an update?

@lavamind I made a fairly simple workflow for this:

workflow-nodes-2026-11-19.json (4.4 KB)