summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Martin Pitt [Fri, 22 Apr 2022 12:10:36 +0000 (14:10 +0200)]
Move to official Grafana container image on docker.io
The bitnami image ceased to exist.
Martin Pitt [Fri, 22 Apr 2022 11:57:35 +0000 (13:57 +0200)]
Move to different nginx container
quay.io/bitnami/nginx does not exist any more. This one is maintained
well.
Also expose the httpd port to the host, for easier debugging.
Martin Pitt [Fri, 9 Jul 2021 07:32:39 +0000 (09:32 +0200)]
Reorganize configs
- Rename directories like the volumes.
- Add Grafana dashboard provisioning, and split off config. It was weird
to mount the same directory under two different paths.
- Now arbitrary dashboards can be put into
grafana-provisioning/dashboards/*.json and they will be automatically
deployed.
Martin Pitt [Wed, 16 Jun 2021 05:23:14 +0000 (07:23 +0200)]
Move Grafana config to directory volume
This more closely resembles what one would do on Kubernetes with a
`ConfigMap`.
Martin Pitt [Wed, 16 Jun 2021 05:05:50 +0000 (07:05 +0200)]
Grafana: Enable anonymous access
Martin Pitt [Wed, 26 May 2021 08:13:45 +0000 (10:13 +0200)]
Pull prometheus image from quay.io
To avoid docker.io pull limits
Martin Pitt [Wed, 26 May 2021 06:52:23 +0000 (08:52 +0200)]
Move Prometheus config to directory volume
This more closely resembles what one would do on Kubernetes with a
`ConfigMap`.
Martin Pitt [Mon, 24 May 2021 10:03:35 +0000 (12:03 +0200)]
Add histrogram example
Martin Pitt [Fri, 21 May 2021 11:26:23 +0000 (13:26 +0200)]
Add script to generate time series
The text import format supports specifying a time stamp [1]. Use this to
generate a time series for thing_count and thing_failures.
./things-series > http/metrics
podman pod rm -f learn-metrics || true
podman play kube learn-metrics.yaml
First non-trivial query from "indefinitely long lookback" counters: "How
many tests happened in the last 5 minutes?
rate(thing_count{job="host-http"}[5m]) * 5 * 60
The re-scaling is because the natural unit is "1/s" and we want the unit
to be the same as the original counter.
[1] https://prometheus.io/docs/instrumenting/exposition_formats/
Martin Pitt [Fri, 21 May 2021 08:40:32 +0000 (10:40 +0200)]
Add custom prometheus data source from host
Export http/metrics into prometheus via an intermediate web server in
the pod, and configuring it as an additional source.
You can manually provide metrics by copying http/metrics.template to
http/metrics and edit the numbers while prometheus is running.
Martin Pitt [Fri, 21 May 2021 06:56:59 +0000 (08:56 +0200)]
Replace launch script with kubernetes resource
This can be deployed locally with
podman play kube learn-metrics.yaml
Generated with
podman generate kube learn-metrics > learn-metrics.yaml
and heavy editing/manual cleanup.
Martin Pitt [Fri, 21 May 2021 06:34:06 +0000 (08:34 +0200)]
Pre-configure prometheus data source in Grafana
This mysteriously fixes the data source test "not found" error.
Martin Pitt [Fri, 21 May 2021 06:08:03 +0000 (08:08 +0200)]
Run unconfigured grafana and prometheus containers
Start with `./launch.sh`, tear down with `podman pod rm -f learn-metrics`.
Prometheus explorer at http://localhost:3001 (to avoid port conflict
with Cockpit). It can list or graph simple queries like
`go_memstats_heap_objects` or `go_memstats_heap_objects[5min]`.
Metrics source at <http://localhost:3001/metrics>
<http://metrics-prometheus:9090/metrics> (in grafana container).
Grafana UI at <http://localhost:3000>, log in as admin:foobar. Manually
add Prometheus data source at <http://metrics-prometheus:9090/metrics>.
This fails with "HTTP Error Not Found" with "Server" access (which is
usually what you want -- "Browser" access fails on CORS).