Compatibility with Grafana 8/9

[amotl]

Hi there,

first of all, I want to thank @jm66, @MarcoPignati, @ShagoY, @andsens and @wols for reporting problems with Grafana 8 and for helping to resolve the issues in one way or another (see #103 and #109).

Now, I wanted take the chance to ask which data sources (InfluxDB, Elasticsearch, PostgreSQL, MySQL, ...) all of you are using? It looks like others are also reporting problems using the original Grafana Worldmap panel, which might depend on the type of data source they are using.

• Data Error - TypeError: Cannot read property 'length' of undefined grafana/worldmap-panel#319
• Worldmap panel and Grafana 8 grafana/worldmap-panel#323
• https://community.grafana.com/t/grafana-8-0-incompatibilities/48840
• https://community.grafana.com/t/worldmap-plugin-copy-panel-not-working/48900

Mitigations for those problem reports apparently have been added through either a patch for Worldmap (grafana#318) or a patch to Grafana itself (grafana/grafana#36737). The latter is apparently already available through Grafana 8.0.6. Thank you for your efforts on this, @hugohaggmark, @mckn, @ryantxu and @torkelo.

So, this is currently a placeholder issue in order to investigate and discuss how those improvements might also have to be applied by users of the Panodata Map Panel fork in any way.

With kind regards,
Andreas.

[mckn]

The patch that includes the transformation are also available in 7.5.10 for those still running Grafana 7.

[jm66]

Hello, we are using PostgreSQL on Grafana v8.0.6. The original Grafana Worldmap panel is installed (0.3.3) but not being used 😅

[duylong]

Hi,
Grafana Worldmap panel (0.3.3) no works with Elasticsearch 7 and Grafana 8.06.

[MarcoPignati]

We are using Flux to call postgresql (with sql.from)

[BinaryPaean]

I'm seeing various related errors attempting to get data from PostgreSQL with Grafana v8.0.6 and current (as of time of writing) Worldmap Panel.

[amotl]

Hi there,

we just released panodata-map-panel-0.16.0 and will be happy to receive any feedback on this with respect to compatibility with Grafana 8.

With kind regards,
Andreas.

[duylong]

Hi there,

we just released panodata-map-panel-0.16.0 and will be happy to receive any feedback on this with respect to compatibility with Grafana 8.

With kind regards,
Andreas.

I have tested with this release. Same problem but I noticed this error "GET https://XXX/grafana/public/plugins/grafana-worldmap-panel/module.js?_cache=1626873546956 net::ERR_UNEXPECTED_PROXY_AUTH". When I go to the link, I have the "failed to get the user" error on the page. No problem with the other panels.

Grafana log:

juil. 21 15:22:02 XXX grafana-server[19642]: t=2021-07-21T15:22:02+0200 lvl=eror msg="Failed to login" logger=auth.proxy username=(null) message="failed to get the user" error="did not find a user" ignoreCache=false
juil. 21 15:22:02 XXX grafana-server[19642]: t=2021-07-21T15:22:02+0200 lvl=eror msg="failed to get the user" logger=context error="did not find a user"
juil. 21 15:22:02 XXX grafana-server[19642]: t=2021-07-21T15:22:02+0200 lvl=info msg="Request Completed" logger=context userId=0 orgId=0 uname= method=GET path=/public/plugins/grafana-worldmap-panel/images/worldmap_logo.svg staatus=407 remote_addr=XXX time_ms=16 size=716 referer=https://XXX/grafana/plugins

[amotl]

Hi @duylong,

thanks for your report. Requests to /public/plugins/grafana-worldmap-panel might indicate that those are not originating from a panel instance of this fork.

Can you make sure to uninstall all other map panels and check the Grafana "plugins" folder for any leftovers beforehand? Maybe also uninstall/remove and reinstall panodata-map-panel just to be more safe than sorry?

With kind regards,
Andreas.

[wols]

Is this a register problem by you?

# date -u --iso=m
2021-07-22T21:09+00:00
# grafana-cli plugins install panodata-map-panel
failed to find plugin 'panodata-map-panel' in plugin repository. Please check if plugin ID is correct

Error: plugin not found

[amotl]

Hi @wols,

this panel never has been available in the Grafana Plugin Repository yet. We are planning to do it though (#112).

With kind regards,
Andreas.

[raphaelrrl]

DataError: Format mismatch: table data can not be interpreted as timeseries format

error adding 2 locations in Json format Grafana 8.2

QUERY A

QUERY B

[mckn]

DataError: Format mismatch: table data can not be interpreted as timeseries format

error adding 2 locations in Json format Grafana 8.2

QUERY A

QUERY B

There is a transformation you can add that will map the new data source structure to the old time series format.

[matschaffer]

I was trying to get https://github.com/panodata/panodata-map-panel/releases/tag/0.16.0 running on 8.2.5 today, but no success. Seems like grafana is just refusing to load the plugin.

[d-roet]

I was trying to get https://github.com/panodata/panodata-map-panel/releases/tag/0.16.0 running on 8.2.5 today, but no success. Seems like grafana is just refusing to load the plugin.

@matschaffer I think that by default Grafana treats unsigned plugins this way. Currently I am running my Grafana server (v8.3.3) in development mode instead of the default production mode, you can do this by altering /etc/grafana/grafana.ini like and then restarting.

# possible values : production, development
# TEMPORARY TO ALLOW UNSIGNED PLUGINS!
;app_mode = production
app_mode = development

After that the grafana.log file should show entries like:

t=... lvl=warn msg="Permitting unsigned plugin. This is not recommended"

But the panodata-map-panel does come through in our instance.

[usommerl]

@d-roet @matschaffer FYI: You can load unsigned plugins in production mode when you whitelist these plugins in your Grafana configuration. See the allow_loading_unsigned_plugins configuration option in the plugins section. You can also use the GF_PLUGINS_ALLOW_LOADING_UNSIGNED_PLUGINS environment variable to configure the same setting via the environment at startup.

[amotl]

Hi again,

we recently upgraded to Grafana 9 on one of our servers, and are able to confirm that Panodata Map Panel 0.16.0 is working well ✅.

That being said, we would still have to improve some details on the editor side, as reported by @thepilot21 at GH-126. Further, the DataError: Format mismatch error observed on specific data sources, and reported by @raphaelrrl and @mckn, is being tracked at GH-123.

If you are observing other issues we are not aware of, please open a separate issue.

With kind regards, and a happy new year,
Andreas.