azure-functions-host: Azure function time trigger not resuming unless I visit the azure portal

Investigative information

Timestamp: 2018-02-13T15:35:00.0076347+00:00 (successfull run) / 2018-02-14T08:02:05.0220800+00:00 (next successfull run … no activity in between, this is supposed to be an every 5 mins trigger)
Function App version (1.0 or 2.0-beta): Runtime version: 1.0.11535.0 (~1)
Function App name: FdriverFunctionsDemo
Function name(s) (as appropriate): PullXmlFilesTimer
Invocation ID: 5a170853-ea00-456f-9ae0-b151454f2048 / c16cf1c2-cdf2-49b1-92ee-209806694ce1
Region: West-Europe

Repro steps

I have a timer triggered azure function with schedule = “0 */5 * * * *”, the function is supposed to run every 5 minutes. It stops after a few invocations… I don’t see any error in the various available logs. …

Expected behavior

The function should run every 5 minutes

Actual behavior

The function runs a few times and then stops. There are no new invocations unless I manually go the azure portal website and look at the azure function… then it starts for a few rounds and stops again.

Known workarounds

N/A

Related information

Programming language used : C#
Links to source : Not publically available
Bindings used

    {
      "type": "timerTrigger",
      "schedule": "0 */5 * * * *",
      "useMonitor": true,
      "runOnStartup": false,
      "name": "myTimer"
    }
  ]

About this issue

Original URL
State: closed
Created 6 years ago
Reactions: 4
Comments: 38 (16 by maintainers)

Most upvoted comments

I have filed an issue here to make it so we are a little more lenient with that host.json file.

@vludax I suspect that our internal keep-alive logic for HttpTriggers is a little bit different than our keep-alive logic for TimerTriggers. At an HTTP request every 5 minutes, you function application was always kept alive, so even though the trigger was never synchronized, since your application was never recycled, that worked out fine. I recommend manually synchronizing once you remove the comments from host.json, because until your http trigger is synchronized, if your application is ever recycled, it may stop triggering on HTTP requests until the application is loaded up again.

I am going to close this issue for now, as the original post has been addressed, and we seem to have resolved the other issues that have come up here. If anyone has any issues with triggers not firing, your host.json is valid and the function was deployed using one of the supported methods detailed in this thread, please open up a new issue and we will investigate it there.

ConnorMcMahon on Mar 6, 2018