sitespeedio / grafana-bootstrap-docker

You can find the example dashboards in sitespeed.io in https://github.com/sitespeedio/sitespeed.io/tree/main/docker/grafana/provisioning/dashboards

grafana-bootstrap-docker's Issues

Graphite basicauth not actually authenticated

I noticed that the basic auth credentials in /datasources/graphite/graphite.json (guest/guest) do not actually get authenticated.
I know that these credentials (guest/guest from .htpasswd) are used for the graphite (admin) dashboard, but how does Graphite as a datasource actually use authentication?

please create tags

Hi, I believe this repo is the home of the grafana dashboards, which I pull directly from this repo when provisioning the dashboards (I use my own process to get sitespeedio running, as I run it on k8s). I have been using the main branch, but I noticed today that it was updated with changes for 17.0.0.

Can you please start creating tags or use release branches so that I can pin the versions of the dashboards down?

At the moment, I've switched to using commit references, but it's hard to tell what version of sitespeedio the dashboards are for from that.

SiteSummary Dashboard composes wrong query

The prepared Dashboards are great to integrate sitespeed with Grafana.

For example https://github.com/sitespeedio/grafana-bootstrap-docker/blob/main/dashboards/graphite/SiteSummary.json gives a nice overview of the measured metrics.

These are the queries composed for the first panel "First Visual Change"

sitespeed_io.pageSummary.summary....browsertime.visualMetrics.FirstVisualChange.median
sitespeed_io.pageSummary.summary....browsertime.visualMetrics.FirstVisualChange.max
sitespeed_io.pageSummary.summary....browsertime.visualMetrics.FirstVisualChange.min

This is the parametrized composition:

The issue is (I guess) the combination pageSummary.summary which doesn't exist. It is either pageSummary or summary.

Any objections to change that globally?

I see that in the global Dashboard config this config exists:

So, based on that it should work. However the config for the dashboard overwrites this:

Third party CPU usage data not visible in Dashboard

Hello,

I enabled --thirdParty.cpu when running sitespeed.io, but on the dashboard, in the CPU time spent per tool/domain - enable using --thirdParty.cpu [Chrome only] section, I only have empty data, even if I have (sporadic) data in graphite.

Target aliasByNode($base.$path.$testname.pageSummary.$group.$page.$browser.$connectivity.thirdparty.tool.*.cpu, 10) looks okay though 🤔

Since Grafana 5.0 this project seems to be redundant

https://grafana.com/docs/administration/provisioning/
Grafana 5.0 added functionality to provision dashboards and data sources

I was about to use this project by achieved very nice integration with dashboards being added automatically without rebuild or restart of containers using provisioning provided by grafana.

here is the reference project https://github.com/vegasbrianc/prometheus/tree/master/grafana/provisioning

Use variables to configure node index parameter in aliasByNode

At the moment, all the metric queries use hardcoded node index number to invokealiasByNode function, e.g.

aliasByNode(sitespeed_io.$path.summary.$group.$browser.$connectivity.browsertime.pageTimings.*.median, 8);

Here is an example in the code base,

grafana-bootstrap-docker/dashboards/pageSummary.json

Line 627 in b04d279

"target": "aliasByNode(sitespeed_io.$path.pageSummary.$domain.$page.$browser.$connectivity.browsertime.statistics.timings.pageTimings.*.median, 11)",

This makes it very tedious to add new dimensions (e.g. such as this requirement sitespeedio/sitespeed.io#1327).

Consider using variables, e.g.

aliasByNode(sitespeed_io.$path.summary.$group.$browser.$connectivity.browsertime.pageTimings.*.median, $aliasNodeDepth);

"bulk" data messes dashboard

Hello,

New version of sitespeed.io (16.10) seems to break current dashboard: it adds a new "bulk" data level graphite, which renders current queries unusable.
Is there a way to fix it?

template vars import doesn't work for constant `base`

After import dashboards the __input constant base is set to base instead of sitespeed_io (defined as default? in pageSummary.__inputs)

Is there any API change in grafana 4? Any ideas?

ResultURL variable in new dashboards

The variables are pointing to: https://s3.amazonaws.com/results.sitespeed.io
This is an issue for anyone who is running their instances locally or even storing elsewhere.

grafana-bootstrap-docker/dashboards/graphite/PageMetricsMobile.json

Line 22 in e63cd58

"value": "https://s3.amazonaws.com/results.sitespeed.io",

In addition to this bug. The dashboard will display as shown in the image.

Graphe the timing spent on each 3rd party server

On the old demo dashboard, we had a nice graph showing the slowest assets, over time.

It was nice because it could be used to look at the availability and performance of the 3rd party (and some slow 1st party resource).
We could produce at least for the public demo a graph showing the response time per domain (it will probably store less data). It wont tell much about 1st party resources having troubles, but in my case it will be enough to monitor how 3rd parties scripts are performing.

Dashboard for version 29.0.0 adds unknwon JSON datasource

Hello @soulgalore ,

You recently updated page metrics' dashboard's JSON, but you added a new JSON API Datasource labeled DS_JSON_API.
But there seem to be no clue on how it is used.
Is a mistake or it misses associated documentation? 🙂

I'm available for more details is needed :)

Easy way to translate Graphite dashboards to InfluxDB?

We recently adopted InfluxDB, and I was sad to see there was only the PageSummary dashboard under the examples: https://github.com/sitespeedio/grafana-bootstrap-docker/tree/master/dashboards

Is there an easy way to translate? A fork or other examples out there?

Thanks!