This repo contains some basic examples for monitoring tasks around Opencast, using the Grafana ecosystem.
The file opencast-metrics-dashboard.json.j2 contains a basic Grafana dashboard for Opencast.
The dashboard uses the label_replace() function a lot, e.g. in the form of label_replace(your_metric, \"instance\", \"$1\", \"instance\", \"(.*).example.com.*\"). Those variables have been templated here like this:
| variable name | example value |
|---|---|
your_domain_regex |
(.*).example.com.* |
your_worker_regex |
https://oc-worker[0-9]{2}.example.com |
your_admin_regex |
oc-admin.example.com.de.* |
your_presentation_regex |
oc-presentation.example.com.de.* |
The file opencast_promtail_config.yml.j2 contains a basic logging config for promtail with Opencast.
It can directly be fed to the ansible role monitoring_promtail via the variable promtail_config_template.
The following variables are somehow expected by this template, the default values refer to the ones set by the ansible role.
| variable name | (default-)value |
|---|---|
inventory_hostname |
required, the current host, e.g. oc-admin.example.com |
loki_host |
required |
loki_http_listen_port |
required |
promtail_use_tls |
false |
promtail_http_listen_port |
9080 |
promtail_grpc_listen_port |
0 |
use_basic_auth |
false |
basic_auth_user |
required if use_basic_auth is true |
basic_auth_password |
required if use_basic_auth is true |
promtail_job_journal |
false |
The file opencast-prometheus.yml contains an example for prometheus scrape jobs that expands on the example from the docs for a multi-node cluster.
Specifically, it splits up the metrics from the opencast endpoint into two jobs:
- Opencast-specific metrics that are only scraped from the admin node
- JVM metrics that are scraped from all nodes