Number of messages used from iot edge

Question

Goodmorning,

I have seen on iot hub's metric some strange values..  I have actually one edge ( raspberry) power on with only two modules running, edgeAgent and edgeHub. Other modules are currently stopped.

- Max Successful twin reads from devices --> one per hour

- Avg Response size of twin reads from devices  --> about 9.3 KB  one time each hour

- Total number of messages used --> it increased by about 30 messages every hour, probably because one twin read it will be count with block's size of 512 bytes

What do these updates actually do every hour ?

Why does my number of messages increase for automatic device operations? Can i reduce the frequency ?

So, i looked into iot edge logs, and i saw on edgeHub'logs :

2019-04-09 02:10:20.729 +00:00 [INF] - Starting compaction of stores
2019-04-09 02:10:20.730 +00:00 [INF] - Starting compaction of store default
2019-04-09 02:10:20.730 +00:00 [INF] - Starting compaction of store checkpoints
2019-04-09 02:10:20.730 +00:00 [INF] - Starting compaction of store twins
2019-04-09 02:10:20.731 +00:00 [INF] - Starting compaction of store messages
2019-04-09 02:10:20.731 +00:00 [INF] - Starting compaction of store sessions
2019-04-09 02:10:20.731 +00:00 [INF] - Starting compaction of store DeviceScopeCache
2019-04-09 02:10:20.757 +00:00 [INF] - Starting compaction of store iothub
2019-04-09 02:10:29.760 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 02:10:30.114 +00:00 [INF] - Starting refresh of device scope identities cache
2019-04-09 03:10:29.749 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 03:10:30.763 +00:00 [INF] - Starting refresh of device scope identities cache
2019-04-09 04:10:20.721 +00:00 [INF] - Starting compaction of stores
2019-04-09 04:10:20.721 +00:00 [INF] - Starting compaction of store default
2019-04-09 04:10:20.722 +00:00 [INF] - Starting compaction of store checkpoints
2019-04-09 04:10:20.722 +00:00 [INF] - Starting compaction of store twins
2019-04-09 04:10:20.722 +00:00 [INF] - Starting compaction of store messages
2019-04-09 04:10:20.723 +00:00 [INF] - Starting compaction of store sessions
2019-04-09 04:10:20.723 +00:00 [INF] - Starting compaction of store DeviceScopeCache
2019-04-09 04:10:20.743 +00:00 [INF] - Starting compaction of store iothub
2019-04-09 04:10:29.749 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 04:10:31.471 +00:00 [INF] - Starting refresh of device scope identities cache
2019-04-09 05:10:29.751 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 05:10:32.120 +00:00 [INF] - Starting refresh of device scope identities cache
2019-04-09 06:10:20.719 +00:00 [INF] - Starting compaction of stores
2019-04-09 06:10:20.719 +00:00 [INF] - Starting compaction of store default
2019-04-09 06:10:20.814 +00:00 [INF] - Starting compaction of store checkpoints
2019-04-09 06:10:20.815 +00:00 [INF] - Starting compaction of store twins
2019-04-09 06:10:20.815 +00:00 [INF] - Starting compaction of store messages
2019-04-09 06:10:20.815 +00:00 [INF] - Starting compaction of store sessions
2019-04-09 06:10:20.815 +00:00 [INF] - Starting compaction of store DeviceScopeCache
2019-04-09 06:10:20.834 +00:00 [INF] - Starting compaction of store iothub
2019-04-09 06:10:29.755 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 06:10:32.830 +00:00 [INF] - Starting refresh of device scope identities cache
2019-04-09 07:10:29.749 +00:00 [INF] - Reauthenticating connected clients
2019-04-09 07:10:33.502 +00:00 [INF] - Starting refresh of device scope identities cache

Answer 1

Thanks for your feedback Greta,

We are looking at this internally. Please bear with me.

Answer 2

Hello Greta,

The Edge Agent periodically syncs the deployment to prevent us from ever "bricking" the device. The default period for this sync is 1 hour, which is what you are seeing here.

This period can be adjusted by setting the "ConfigRefreshFrequencySecs" environment variable in the Edge Agent.

See: https://github.com/Azure/iotedge/blob/f43ccb30b58dfba87a479a8afd0b7db21834680b/edge-agent/src/Microsoft.Azure.Devices.Edge.Agent.Service/appsettings_agent.json

Answer 3

Thanks a lot!

So, is it only used for sync deployment?

How can i change it for edgeAgent? I think i should edit deployment.json, adding env section for edgeModule.

If i will change the ConfigRefreshFrequencySecs to 6 hours, what happens if i  I create a new pipeline deployment in the middle of the six hours? It will immediately applied or after 6 hours?

 

Answer 4

Good morning,

I tried to put ConfigRefreshFrequencySecs in different places in my deployment.json and also inside the config.yaml, but it seems edgeHub doesn't use the new value.

In deployment.json i tried to insert it both for edgeAgent and edgeHub. I noticed  only  logLevel can use a new value.

"systemModules": {
          "edgeAgent": {
            "type": "docker",
            "env": {
              "ConfigRefreshFrequencySecs": {
                "value": 1800
              }
            },
            "settings": {
              "image": "mcr.microsoft.com/azureiotedge-agent:1.0",
              "createOptions": "{}"
            }
          },
          "edgeHub": {
            "type": "docker",
            "status": "running",
            "restartPolicy": "always",
            "env": {
              "ConfigRefreshFrequencySecs": {
                "value": 800
              },
              "RuntimeLogLevel": {
                "value": "debug"
              }
            },
            "settings": {
              "image": "mcr.microsoft.com/azureiotedge-hub:1.0",
              "createOptions": "{}"
            }
          }
        }

And i tried also to put it inside config.yaml, i noticed that is taked it but with first lowercase char. This with the command line " docker exec -it edgeHub env "

agent:
  name: "edgeAgent"
  type: "docker"

  config:
    image: "mcr.microsoft.com/azureiotedge-agent:1.0"
    env :  { ConfigRefreshFrequencySecs : 800 }
auth: {}

Where  i should  insert it ?

Answer 5

Hello Greta,

You may be running into a known issue with triggering an update of the Edge Agent module specifically. Currently only changing the "image" field triggers a redeployment of the Edge Agent module.

The env setting you have should work. I would update the "image" field to mcr.microsoft.com/azureiotedge-agent:1.0.7 to trigger the redeployment.

Let us know the result?

Thanks!

Answer 6

I have tried, but nothing to do.... any news ?

Answer 7

Hi Greta,

This may be a bug in the SDK and we need to validate. Can I kindly ask you to file a new issue on the SDK repo?

https://github.com/Azure/iotedge/issues

Thank you so much. Please inform here the link to the issue created so others can follow as well :).

Answer 8

Of course,

here the issue opened https://github.com/Azure/iotedge/issues/1462 

Answer 9

Is there any documentation about edgeAgent's setting  from  here ?