periodic API errors with Azure via HTTP

[sphen]

Love the new features of 6.2, thank you!

I am testing out the new Azure via HTTP template and I am having an issue.

Somewhat randomly on a discovered virtual machine, the script that calls the api and parses values has an issue where it cannot parse certain values. For example:

"errors": "Failed to receive data:\nPercentage%20CPU,Network%20In,Network%20Out ,Disk%20Read%20Bytes,Disk%20Write%20Bytes,Disk%20R ead%20Operations%2FSec,Disk%20Write%20Operations%2 FSec,CPU%20Credits%20Remaining,CPU%20Credits%20Con sumed,Data%20Disk%20Read%20Bytes%2Fsec,Data%20Disk %20Write%20Bytes%2Fsec,Data%20Disk%20Read%20Operat ions%2FSec,Data%20Disk%20Write%20Operations%2FSec, Data%20Disk%20Queue%20Depth,Data%20Disk%20Bandwidt h%20Consumed%20Percentage,Data%20Disk%20IOPS%20Con sumed%20Percentage,Data%20Disk%20Target%20Bandwidt h,Data%20Disk%20Target%20IOPS,Data%20Disk%20Max%20 Burst%20Bandwidth,Data%20Disk%20Max%20Burst%20IOPS : TypeError: cannot read property 'data' of undefined",

But then later it will be fine. As far as i can tell my server is not overloaded, the timeouts are standard in the inherited macros, nothing in azure should be limiting our reporting.

When the script pulls the json is there a way to get the raw json so we can see if it is malformed or truncated perhaps?

Has anyone else tried this out and had issues? Interestingly enough our mysql instance does not have API errors but all of our VMs seem to (but because of this error i disabled all but 1 for now).

Thanks for any direction.

[Markku]

Looks like the template has logging in level 4, so increase the logging level in the server and see the server log for the information. Or, edit the scripts so that the logging occurs already in level 3.

Markku

[sphen]

good idea - dropping the logged events to 3 - will see. do many people use this template yet? prob not?

[sphen]

so - seems when this comes up - yeah the api is returning not all data - why? i have no clue. for example, it will return 9 metrics when its an issue and 19 when its ok.

no one else having issues with this or using the https api?

[sphen]

nvm i need to look more - seems the script breaks the requests up into ~ 20 per request. im not a programmer so its a bit difficult to weed through this

[RainCap]

Same problem here, for example I can't get available memory , was able to get it once, now it's not supported. I wonder if it's Azure problem. Ideas?

[sphen]

seems the api calls dont always return all data requested. i was looking more specifically for cpu credits etc, so i have removed many of the metrics form the script that i dont really need for now. this has reduced the amount of api errors dramatically. i have also disabled the api triggers i am curious what the contributor of this code has to say on the topic?