Receive an error when create an alter

[yuanpenc]

What happened:

When I was trying to create an alter and test rules, there seemed to be a bug. The request and response is as below.

  "firing": true,
  "state": "pending",
  "conditionEvals": " = true",
  "timeMs": "179.965ms",
  "error": "tsdb.HandleRequest() response error &{POST http://proxy.druid.ke.com/druid/v2 giving up after 6 attempt(s) POST http://proxy.druid.ke.com/druid/v2 giving up after 6 attempt(s) A <nil> [] [] 0xc001913e30}",
  "logs": [
    {
      "message": "Condition[0]: Query",
      "data": {
        "from": 1622531328676,
        "queries": [
          {
            "refId": "A",
            "model": {
              "builder": {
                "aggregations": [
                  {
                    "name": "all_cnt",
                    "type": "count"
                  }
                ],
                "dataSource": {
                  "name": "commerce_adx_digv_empty_log",
                  "type": "table"
                },
                "filter": {
                  "dimension": "ip",
                  "type": "selector",
                  "value": "10.26.12.246"
                },
                "granularity": "minute",
                "intervals": {
                  "intervals": [
                    "${__from:date:iso}/${__to:date:iso}"
                  ],
                  "type": "intervals"
                },
                "postAggregations": [],
                "queryType": "timeseries",
                "virtualColumns": []
              },
              "expr": "{\"builder\":{\"queryType\":\"timeseries\",\"dataSource\":{\"name\":\"commerce_adx_digv_empty_log\",\"type\":\"table\"},\"intervals\":{\"type\":\"intervals\",\"intervals\":[\"${__from:date:iso}/${__to:date:iso}\"]},\"granularity\":\"minute\",\"filter\":{\"dimension\":\"ip\",\"type\":\"selector\",\"value\":\"10.26.12.246\"},\"aggregations\":[{\"name\":\"all_cnt\",\"type\":\"count\"}],\"postAggregations\":[],\"virtualColumns\":[]},\"settings\":{}}",
              "hide": true,
              "refId": "A",
              "settings": {}
            },
            "datasource": {
              "id": 18,
              "name": "Druid-server-alert"
            },
            "maxDataPoints": 0,
            "intervalMs": 0
          }
        ],
        "to": 1622531628676
      }
    }
  ]
}

As you can see, the parameter "intervals" seems not work properly. At lease, I think, It should not be something like "${__from:date:iso}/${__to:date:iso}".

However, When I specified a certain time range like the pic below, the problem was solved.

What you expected to happen:
The alter works successfully without specifying the intervals.

For an instance, when I use official grafana-prometheus. The request of an alter is like this:

The interval is "" (nothing here, only "from/to" specify the time range )

How to reproduce it:
Build an timeseries query without specifying the intervals. Then, create an alter and test rule. The error message will come out.

Environment:

• Grafana version: v7.3.5
• Data source type: druid
• OS Grafana is installed on: Linux
• User OS & Browser: Chrome

[jbguerraz]

Hello @yuanpenc
Thank you for this issue! :)
Yes this issue is known, has been discussed here: https://grafadruid.slack.com/archives/C01EEPVTALW/p1621957357043800
We should replace the __from and __to variables from the backend, using those from and to fields from the alert query, and then apply the formaters over them (like date:iso or so: https://grafana.com/docs/grafana/latest/variables/variable-types/global-variables/#__from-and-__to ).
If you feel it, feel free to join us on slack and maybe contribute this feature ?
For short-term workaround you could use something relative to now to select the data (like now - 5 minutes or so) for your alerts.

[yuanpenc]

Hello @yuanpenc
Thank you for this issue! :)
Yes this issue is known, has been discussed here: https://grafadruid.slack.com/archives/C01EEPVTALW/p1621957357043800
We should replace the __from and __to variables from the backend, using those from and to fields from the alert query, and then apply the formaters over them (like date:iso or so: https://grafana.com/docs/grafana/latest/variables/variable-types/global-variables/#__from-and-__to ).
If you feel it, feel free to join us on slack and maybe contribute this feature ?
For short-term workaround you could use something relative to now to select the data (like now - 5 minutes or so) for your alerts.

Of course! I am 100% willing to contribute this feature. But the thing is I don't have too much experience in Typescript project. So, Could you please add some instructions of how to compile and zip this project in the readme? Thank you for you help.

[jbguerraz]

Hello @yuanpenc
Awesome to read! This is more a contribution on the golang side of it. I'm fully available to help you with it, feel free to ping me on https://join.slack.com/t/grafadruid/shared_invite/zt-ryrcolyp-GQCiKcOF0WhnK~iqpkrqLA

[saketbairoliya2]

Do we have any update on this? @jbguerraz / @yuanpenc

[jbguerraz]

nope @saketbairoliya2
If you're willing to work on it, feel free to ping me, I'll be happy to help!